• 1pea TheWorkla.nk / I I I • I I Proceedings of the Third International Urban Research Symposium.held in Brasilia April 2005 Land and Urban Policies for Poverty Reduction Volume 2 EDITED BY Mila Freire Christine Kessides Ricardo Lima Jose Aroudo Mota Dean Cira Diana Motta · Bruce Ferguson . •• The World Bank 1pea . Brasilia, 2007 Federal Governement Ministry of Planning, Budget and Management Minister - Paulo Bernardo Silva Executive Secretary - Joao Bernardo de Azevedo Bringel The World Bank • The World Bank Group is the world's largest source of development assistance. Its mission is to fight poverty and improve the living standards 1pea Institute for Applied Economic Research of people in the developing world. It is a development bank, providing Joans, policy advice, technical assistance, and knowledge sharing services lpea is a public fundation linked to the Brasilian Ministry of Planning, to low-and-middle income countries to "reduce poverty. The Bank Budget and Management. It provides technical and institucional support promotes growth to create jobs and to empower poor people to take to governmental actions for public policy making and development 3dvantage of these opportunities. It uses its financial resources, trained programs in Brazil. lpea makes vailable to society the researches and ;taff, and extensive knowledge base to help each developing country studies carried out by its experts. onto a path of stable, sustainable, and equitable growth in the fight against poverty. The World Bank group has 184 member countries. President Luiz Henrique Proen~a Soares Vice President Latin America Pamela Cox Board of Directors l\nna Maria T. Medeiros Peliano Cinara Maria F. de Lima Director Brazil Department Joao Alberto De Negri John Briscoe Jose Aroudo Mota Paulo Mansur Levy Urban Sector Manager Renato L6es Moreira John Henry Stein Chief of Staff Persia Marco Antonio Davison Sector Leader SDN Jennifer Sara Head of the Communication Departament Murilo Lobo Senior Adviser Urban Mila Freire For more information see www.worldbank.org/data/. Ombudsman: http://www.ipea.gov.br/ouvidoria URL: http://www.ipea.gov.br Land and Urban Policies for Poverty Reduction EDITED BY Mila Freire Christine Kessides Ricardo Lima Jose Aroudo Mota Dean Cira Diana Motta Bruce Ferguson © 2007 Institute for Applied Economic Research - Ipea and The International Bank for Reconstruction and Development - The World Bank The findings, interpretations and conclusions expresse~ in this publication do not necessarily reflect the · views of the World Bank and the Instituto de Pesquisa Economica Aplicada (IPEA). The World Bank and the IPEA do not guarantee the accuracy of the data included in this work. The reproduction of these texts and the data in contained is allowed, provided the source is cited. Reproductions for commercial purposed are not allowed. The contents of this publication were subject to final editing by the World Bank. Land and urban policies for poverty reduction : proceedings of the third International Urban Research Symposium I edited by Mila Freire ... [et al.]. - Washington, DC: World Bank; Brasilia: Ipea, 2007. 2 v. : ill. Includes bibliographical references. 1. Urban Land. 2.Urban Policy. 3.Poverty. 4.Land Market. 5. Property Rights. 6.Informal Settlement. 7. Urban Planning. 8. Urban Development. 9.Secure Tenu~e. IO.Slums. I. Freire, Mila. II. World Bank. III. Institute for Applied Economic Research. IV. International Urban Research Symposium (3. : 2005 : Brasilia, DF). CDD 711.4 TABLE OF CONTENTS Forward .................................................................................................................................................... 7 Acknowledgments ...................... ~ .............................................................................................................. 9 Contributors ........................................................................................................................................... 11 Glossary of Abbreviations ....................................................................................................................... 17 Preface: Land, Shelter, Transport: The Latin America~ Way .................................................................... 21 Sir Peter Hall General Introduction ............... :.............................................................................................................. 35 Mila Freire, Bruce W Ferguson, Ricardo Lima, Dean Cira and Christine Kessides Introduc;:ao Geral .................................................................................................................................... 47 Mila Freire, Bruce W Ferguson, Ricardo Lima, Dean Cira and Christine Kessides PART IV: TRANSPORT, DENSITY, URBAN PLANNING AND URBAN FORM ....................... 61 Introduction ........................................................................................................................................... 63 Bruce W Ferguson Urban Land Markets and Urban Land Development An Examination of Three Brazilian Cities: Brasilia, Curitiba and Recife ................................................. 67 M. V Serra, David E Dowall, Diana Motta, and Michael Donovan Agglomeration and Urban Productivity: Implications for the Appraisal of Transport Investment .......................................................................................................... 97 Daniel J Graham · Impact of Transport Infrastructure & Services on Urban Poverty And Land Development: A Case Study - Colombo, Sri Lanka ............................................................ 117 Amal S. Kumarage Interpretation of Population Density Gradients: A Brazilian Perspective .......................................................................................................................... 135 Paulo Coelho Avila, and Paul Irving Mandell · In San Jose, Costa Rica, Effective Metropolitan Planning and Selective Infrastructure Investment Can Improve the Quality of Life for the Poor ............................................... 161 Rosendo Pujol PART V: HOUSING MARKETS AND PROGRAMS ................................................................... 189 Introduction ........................................................ ~ ................................................................................ 191 Bruce W Ferguson Seeking Better Policies or Just Giving Up Responsibility? The Decentralization of Argentina's National Housing Fund (FONAVI) .............................................. 197 Cecilia Zanetta The Challenges of State-Municipality Partnerships in Self-Help Housing Programs: Assessment of a Large-Scale Intervention In the State of Sao Paulo - Brazil .......................................................................... .'............................... 213 Carolina Moretti Fonseca, Eduardo Trani, and Tania Wakisaka Affordable Housing Needs Assessment Methodology: The Adaptation of the Florida Model to Brazil ..................................................................................... 239 foseli Macedo, Diep Nguyen, William J O'Dell, Marc T. Smith, M. V Serra, George A. da Guia, Maria da Piedade Morais, Luiz Alexandre R. Paixao, Paulo A. Rego, and Santiago F Varella Squatters no More: Singapore Social Housing ...................................................................................... 269 Belinda Yuen Commons and Anticommons: Role of the State in the Housing Market .. :.......................................... 295 fiemingZhu PART VI: DEVELOPMENT ON THE URBAN FRINGE AND THE CITY CENTER, AND THE ENVIRONMENT ...................................................................................... 311 Introduction ............................................................................................... ,......................................... 313 Bruce W Ferguson Post Modern Urbanization and the Vulnerability of the Poor ................................................................ 317 felena Pantelic, Bogdan Srdanovic, and Marjorie Greene Cities with Suburbs: Evidence from India ............................................................................................ 329 Kala Seetharam Sridhar Urban Sprawl, Land Markets and Environment Degradation In Sao Paulo, Brazil ............................................................................................................................... 357 Haro/do Torres, Humberto Alves and Maria Aparecida de Oliveira Private Residential Investment Growth: Implications on Municipal Revenues and Socio-Economic Indicators: The Case of the Municipality of Pilar ............................................................................................................................. 379 Cynthia Goytia Urban Development in Developing Countries ...................................................................................... 411 Anamaria de Aragao Costa Martins Vacant Areas in Guadalajara, Mexico: A Profile of Property and Owners ............................................. 429 Adriana Fausto Brito FORWARD Within his role to promote and disseminate research and support the Brazilian government in the design, evaluation and follow-up of public policies, the Instituto de Pesquisa EconomicaAplicada, IPEA has partnered with the World Bank in the organization of the Third International Research Urban Symposium. The event which took place in Brasilia in 2005, was supported by several partners including the Swedish International Development Agency (SIDA), the Lincoln Institute of Land Policy, Caixa Economica Federal, o Governo do Distrito Federal - GDF, Cities Alliance, and GTZ. The papers included in this publication represent the contribution of many researchers from all over the world who have been working on the role of urban and land policies in promoting development and alleviating urban poverty. We believe this is an important contribution that will be widely used in the design of urban policies not only in Brazil but in many other countries. Many thanks to all the authors and partners who made this work possible. John Henry Stein Luiz Henrique Proenfa Soares Urban Sector Manager President Latin America & the Caribean Region !pea The World Bank ACKNOWLEDGMENTS This book is the result of a team effort, and as such, it has benefited from an array of invaluable contributions. Our thanks are due to a large number of people. First, the papers' authors - both those who have been selected to be included in this anthology and those who participated in the Symposium with ideas, comments and discussions. They have provided not just material of outstanding technical quality but a remarkable commitment to enriching the debate about the contribution of land and urban policies in the fight against urban poverty. We are fortunate to share this book with these principal authors (in alphabetical order) - Pedro Abramo, Paulo Avila, John Betancur, Adriana Fausto Brito, Robert Buckley, David Dowall, Alain Durand-Lasserve, Rogerio Fernandez, Carolina Fonseca, Cynthia Goytia, Daniel Graham, Sir Peter Hall, Ramin Keivani, Amal Kumarage, Maria Mercedes Maldonado, Anamaria de Aragao Costa Martins, M. M. Mooya, Maria da Piedade Morais, Willam O'Dell, Jelena Pantelic, Glenn Pearce-Oroz, Rosendo Pujol, Markus Ruhling, Remy Sietchiping, Kala Seetharam Sridhar, Haroldo Torres, Belinda Yuen, Cecilia Zanetta, and Jieming Zhu. All authors are researchers or practitioners who have shared i:hrough these papers their insights on the major policy questions that are still with us in terms of how to deal with urban poverty. While this book reflects the authors' own view and not necessarily the view of the World Bank, of IPEA or any of the sponsors, its production was institutionally housed at the World Bank and IPEA for the final editing and the printing and dissemination. We are grateful to the guidance provided by Maryvonne Plessis- Feissard, Director of the Urban Department, World Bank, Makhtar Diop, Director of Infrastructure for Latin America, World Bank, Glauco Arb ix, President ofIPEA, Marcelo Piancastelli, Director of Regional and Urban Studies in IPEA, and Martim Smolka, Director of Latin America Programs in Lincoln Institute of Land Policy. We also appreciated the contribution of David Dowall, Victor Serra, and Marianne Fay who guided us into the next stages. We want to thank the essential sponsorship of our international partners including the Swedish International Development Agency (SIDA) who has been encouraging and supporting the International Urban Symposia since 2002: GTZ, DFID, Lincoln Institute of Land Policy, as well as our Brazilian partners, including CAIXA and ESAF. We recognize the importance of all participants at the Symposium held on April 2005, in Brasilia, and thank Luis Henrique Proern;:a who invited the Bank to hold the Third International Symposium in Brazil. The Brasilia Symposium not only brought together a majority of authors under one roof for three days of candid discussions but also included Brazilian officials, including the Executive Secretary for the Ministry of Cities, Erminia Maricato, the National Secretary of Urban Policies, Raquel Rolnik, and the National Secretary 10 Land and Urban Policies for Porvety Reduction for Housing Development, Jorge Hereda and the Vice-President of CAIXA, Aser Cortines, and Diana Motta, Secretary for Urban Development and Housing in the Federal District. Finally, we are especially grateful to the steering committee that patiently prepared the concept note, received comments from the Bank and partners, made the call for papers, and put in motion the selection process. Special thanks are owed to Mila Freire, Christine Kessides, Ricardo Lima and Dean Cira who helped design the program, select the papers and edit the book. Likewise, we owe appreciation to IPEA's team, represented by Maria da Piedade de Morais, Joao Carlos Ramos Magalhaes, Emmanuel Porto and George Da Guia for their contribution in designing the program, selecting papers, and providing the logistics and organization of the event. Special thanks are owed to Bruce Ferguson who gave us his full commitment and wisdom in reviewing all papers and preparing the introduction for the chapters in the book. We also thank John Henry Stein, Sector Manager - Urban, and Jennifer Sara, Sector Leader for the Finance, Private Sector and Infrastructure Department. We are grateful for their help and excitement of seeing this book coming into life. Final appreciation is due to Laura De Brular for her steady help in keeping records, files, and being the true institutional memory of the International Urban Symposium that the Bank and its partners have been organizing in the last five years. We also thank Julia Canter and Christiana J ohnnides and Zoe Trohanis for their support during the actual Symposium in Brasilia, 2005. Our thanks to all of them. CONTRIBUTORS Alex Abiko is Professor at the Escola Politecnica of the University of Sao Paulo/USP; specialist on urban and housing management; and, is head of the Department of Civil Construction of the Esco la Politecnica of the USP. Pedro Abramo is an economist teaching and conducting research at the Institute of Urban and Regional Planning and Research of the Federal University of Rio de Janeiro, Brazil. He is a member of the editorial staff of Cardenos, an urban and regional research journal published by the IPPUR. Paulo Coelho Avila is a partner of Metroquattro Arquitetura Tecnologia where he has carries out jobs iri urban and regional planning. He has provided consulting services to the Secretary of Urban Development and Housing of the Federal District on the Habitar/Brasil Program, which was sponsored by the Inter- American Development Bank. He is an adjunct faculty in the Department of Architecture and Urbanism at the Instituto de Ensino Superior Planalto in Brasilia. John Jairo Betancur is currently Associate Professor at the University of Illinois at Chicago. A native of Medellin, Colombia, Dr. Betancur has done a large amount of research on squatter settlements-from the study of the processes of settlement through the details of how people go about the enterprise of survival to efforts at regularization. Robert M. Buckley is Urban Advisor in the Finance, Economics and Urban Development Department of the World Bank. He has worked on numerous urban, housing, and housing finance projects for the World Bank over the past twenty years. He is also the author of a book, Housing Finance in Developing Countries, and numerous articles on Housing Finance, Housing Policy and Housing Subsidies. David E. Dowall is director of the Institute of Urban and Regional Development (IURD) at the University of California, Berkeley. Dr. Dowall is also a professor of city and regional planning and former chair of the University's Academic Senate. He is a leading expert on urban economics and infrastructure policy. Alain Durand-Lasserve is Research Director at the CNRS (Centre National de la Recherche Scientifique), France. He is currently attached to the SEDET Research Centre, University Denis Diderot, Paris. He is a member of the Advisory Board of the Global Research Network on Human Settlements (HS-Net) and has published widely on housing and tenure issues for the urban poor. Adriana Fausto Brito is professor in the Federal University of Minas Gerais and Researcher in CEDEPLAR. Dr. Brito has author a large number of publication and research work on demographics in Brazil and elsewhere. Dean Cira is Senior Urban Specialist in the World Bank East Asia and Pacific Region. He has been the manager of many urban development and housing projects and program in Latin America and East Asia and has authored several pieces of analytical work and urban strategies. Dean was part of the steering committee that organized the Third Urban Symposium and managed a wide analytical program on Slum Upgrading in Brazil. Edesio Fernandes is professor at the London School of Economics and author of many articles on Institutions and land policy. His work has had a major impact in looking at law and judicial factors in implementing the Estatuto da Cidade. 12 Land and Urban Policies for Porvety Reduction Bruce Ferguson is an Urban Economist with large experience in land and housing development. His career includes work with the US Housing and Urban Development agency, Inter American Development Ban and the World Bank. Bruce was deeply involved in the editing of the articles included in this publication and authored the introduction material. Mila Freire is Senior Adviser of the Sustainable Development Network in the World Bank and responsible for analytical and policy work on Urban and Municipal Development. Prior to this position, Mila was SectorManager of the Urban Sector in Latin America and Director of the Urban and City Management Program in WBI. She was the co-task leader for the Third SymposiUm together with Christine Kessides. Carolina Moretti Fonseca is master of Public Administration by Escola de Administrac;:ao de Empresas de Sao Paulo da Fundac;:ao Getulio Vargas. She is adviser at CDHU - Companhia de Desenvolvimento Habitacional e Urbano de Sao Paulo. Cynthia Goytia is an Urban Economist and a Local Economic Development Specialist from Argentina. Her professional practice is based on consultancy work on local development issues for municipal and city governments in Latin American countries. She is the Associate Director of the Master's Program on Urban Economics at the Torcuato Di Tella University in Buenos Aires, Argentina. Daniel Graham is a Senior Research Fellow in the Centre for Transport Studies at Imperial College. He received his PhD from the London School of Economics, and his work is in the areas of transport economics and urban and regional economics. Peter Hall (Sir) is the Bartlett Professor of Planning and Regeneration at The Bartlett, University College London and President of the Town and Country Planning Association. He is an internationally renowned authority on the economic, demographic, cultural and management issues that face cities around the globe. Sir Peter has also been for many years a key planning and regeneration adviser to successive governments. He was Special Adviser on Strategic Planning to the British government (1991-94) and a member of the Office of the Deputy Prime Minister's Urban Task Force (1998-1999). Sir Peter is also considered by many to be the father of the industrial enterprise zone concept, adopted by countries worldwide to develop industry in disadvantaged areas. Ramin Keivani is currently Research Coordinator at the Department of Real Estate and Construction - Oxford Brookes University. Dr Keivani is an urban development specialist with a wide range of interests in comparative urban research particularly housing policy in developing and transition economies. He is co- author of a book on housing policy in developing countries and has published a number of papers on globalisation, housing, land markets and urban development in leading international journals. Christine Kessides is lead economist in the Europe and Central Europe Region of the World Bank. Previously she was senior adviser in the urban department and author of the urban strategy published by the Bank in 2000. She was the co-manager of the Third urban symposium together with M. Freire. Amal S. Kumarage is a Professor at the University of Moratuwa, attached to the Division of Transportation Engineering as well as the Chairman of the National Transport Commission in Sri Lanka. He has published over 25 research papers both locally and abroad and has experience working in over 40 major consultancy assignments over the last 10 years. Land and Urban Policies for Porvety Reduction 13 Ricardo Lima is a Brazilian urban policy expert who teaches Economics at the University of Brasilia and represents IBAM (the Brazilian Institute for Municipal Management) in the Federal District. He occupied important positions at the Ministry of Labor and at IPEA, Ministry of Planning, where he was responsible for the area of urban studies for ten years. He is an Electrical Engineer who took his graduate studies in economics at Vanderbilt University and University of California, Berkeley. Joseli Macedo, an architect and urbanist, is an Assistant Professor in the Department of Urban and Regional Planning at the University of Florida. She is a Research Associate with the Shimberg Center for Affordable Housing and she also works as an international development and environmental planning consultant both in Latin America and iri the United States. Marfa Mercedes Maldonado is Associated ~rofessor and researcher at the "Centro lnterdisciplinario de Estudios Regionales....:.. CIDER in the University of Andes. She is also professor of the Lincoln Insitute of Land IPolicy. Her principal area of research is City Law and the process of making law, property regime, legal practices associated to urbanism, land man~gement, and informality. She has been consultant to several projects of land management and as~isted with schemes of using land valuation to finance local infrastructure. Anamaria de Aragao C. Martins works as urban planner and designer for the Government of Distrito Fede- ral, Brazil, having served for the Department of Historical and Artistic Heritage of Brasilia and for the Institute of Urban and Territorial Planning in 1999. At the .present, she works at Distrito Federal State Secretariat for Housing and Urban Development, Brazil, in the revision of Distrito Federal Territorial Masterplan. M~nya Mooya is a lecturer and researcher in the Department of Land Management, Polytechnic of Namibia. He has previo_usly taught in the Department of Land Economy of the Copperbelt University in Zambia. He is a PhD in Real Estate candidate with the University of Pretoria in South Africa. Maria da Piedade Morais is an Economist and is a researcher at the Department of Urb;m and Regional Studies (DIRUR) of the Institute of Applied Economic Research (IPEA) in Brasllia. Her main areas of interest include: Urban Economics, Housing Policy, Urban Indicators, Urban Poverty and Environmental Economics. Jose Aroudo Mota is a Director of Regional and Urban Studies at Ipea, professor of environmental economics at Sustaribility Development Center at the university of Brasilia's and Federal University ofAmazonas. Member of Federal Economic Council. Diana Motta i_s a researcher at the Department of Urban and Regional Studies of the Institute of Applied Economic Research - IPEA in Brasilia. As general urban policy coordinator she led research projects in the fields of urban and housing policy, land use management, urban system, urban upgrading and land market. She was Secretary of Housing and Urban Development of the Federal District of Brazil from 2004 to 2006. · · William O'Dellis ~n Associate Research Professor at the Shimberg Center for Affordable Housing at the University of Florida and mariager of the Florida Housing Data Clearinghouse. He has been involved in local and state government issues in Florida for several years. His local government experience includes housing and capital improvement planning . .· Jelena Pantelic is currently Senior Operation Officer at the Corporate Secretariat's Policy Unit (SECPS) of the World Bank and is in charge of upstream review of strategic and policy documents submitted for the 14 Land and Urban Policies for Porvety Reduction review of the Board of Directors of the World Bank. She has an academic background in urban planning, policy analysis, hazard reduction and architecture and is author of numerous papers related to hazard risk vulnerability, recovery and prevention. Glenn Pearce-Oroz is currently USAID's Sr. Local Governance Advisor in Honduras where he manages decentralization and urban governance activities in 32 cities throughout .the country. After working several years as an urban planner for a municipality in Chile, he joined the World Bank where he worked on natural resource management and social development projects in Brazil and urban development activities in Latin America. Rosendo Pujol created and directs the Research Program on Sustainable Urban Development (ProDUS) of the University of Costa Rica (UCR). He directs and teaches in the Master's Program on Gesti6n Ambiental y Ecoturismo and also teaches undergraduate and graduate courses in the School of Civil Engineering (UCR). He teaches transportation issues for World Bank courses on Urban Management. Carole Rakodi is professor ofinternational urban development in the International Development Department, University of Birmingham, Unit~d Kingdom. She has worked on urban policy and management in Africa since the 1970s and has recently coordinated a study of informal land delivery systems in five African cities. Markus Ruhling after working several years as a consultant for the GTZ, KfW' and World Bank, is now an integrated expert at the Instituto de Investigaci6n y Capacitaci6n Municipal (INICAM) in Peru, in a program sponsored by CIM. His research focus is on development theory, fiscal decentralization, and local government. Morzart Vitor Serra retired from the World Bank in 2004 as a Lead Urban Development Specialist. During his 12-year tenure at the Bank, he led or participated in Bank teams providing policy advice and preparing studies and projects in the fields of municipal development, administration and finance; housing finance and planning; municipal services, public utilities, urban upgrading, and land market; and urban cultural heritage. He has worked in Latin America, Africa, South Asia, and East Asia. Currently, he is a director at Light Servic;:os de Eletricidade in Rio de Janeiro, Brasil. Remy Sietchiping is a Land Tenure Specialist at the UN-HABITAT in Nairobi, Kenya and a a Research Fellow in the School of Social and Environmental Enquire at the University of Melbourne, Australia. He has a Ph.D (Geography) from the University of Melbourne. Dr Sietchiping's work is in the areas of urban planning, slums, land and management and applied GIS. Prior to joining UN-HABITAT, Dr Sietchiping worked as Strategic Resources Analyst for the Victorian State government in Australia, Research Fellow at Deakin University in Australia, GIS Officer at the United Nations Economic Commission for Africa in Addis Ababa (Ethiopia), Geography Lecturer at the University of the West Indies in Jamaica, and Project Manager in an NGO in Cameroon. Kala Seetharam Sridhar is Fellow at, National Institute of Public Finance and Policy (NIPFP) in India. He has written extensively on urban development and the impact of technology in helping the poor. HaroldoTorres is an Economist and a demographer with special training on the issues related to spatial distribution of the population and population and environment. He has also focused his activities on Geographic Information Systems (GIS) applied to social policies. Belinda Yuen is a Chartered Town Planner and Associate Professor in the Department of Real Estate, School of Design and Environment, National University of Singapore. She has published over 80 papers and books Land and Urban Policies for Porvety Reduction 15 on urban planning. She is currently Vice-President of the Singapore Institute of Planners, Honorary Secretary of the ASEAN Association for Planning and Housing (Singapore) and member of several Singapore government committees and civic groups. Cecelia Zanetta is Adjunct Professor at the Department of Geography at the University of Tennessee. She has worked as a consultant to the World Bank and other international financial institutions in several countries in Latin and Central America, including Argentina, Brazil, Chile, Ecuador, El Salvador, Honduras, Paraguay and Peru. Jieming Zhu has been doing research on institutional analysis of urban development in the transitional economy and East Asian cities, and his publications appear in Urban Studies, International Journal of Urban and Regional Research, Urban Affairs Review, Review of Urban and Regional Development Studies, Environment and Planning. He is editorial broad member of Journal of Planning Theory and Practice, Guest Editor for Habitat International. GLOSSARY OF ABBREVIATIONS ABL Public Lighting, Street Sweeping and ~leaning (Argentina) ASEAN Association of East Asian Nations AusAID Australia Assistance International Development AVSI Organizzazione non Governativa di Cooperazione Internazionale BMZ German Ministry of Economic Cooperation BNDES Banco Nacional de Desenvolvimento Econ6mico e Social (Brazil) BACEN Banco Central (Brazil) BIRD Banco Internacional para Reconstruqao e Desenvolvimenfo BPO Business Process Outsourcing BRT Bus Rapit Transit CAFTA Free Trade Agreement of Central America with the United States CANON Municipal Transfer (Peru) CA Cellular Automata CBD Central Business District CB Os Com unity Organizations (South Africa) CCF City Challenge Fund (India) CDHU Companhia de Desenvolvimento Hahitacional e Urbano (Brazil) CEDAE Companhia Estadual de Agua e Esgoto (Brazil) CEDEPLAR Centro de Desenvolvimento e Planejamento Regional da Faculdade de Econbmia da Universidade Federal de Minas Gerais (Brazil) CEF Caixa Econ6mica Federal (Brazil) CEMEX Cement of Mexico COB RAPE Companhia Brasileira de Projetos e Empreendimentos (Brazil) COFOPRI Registration of Informal Urban Property COHRE Center on Housing Rights and Evictions (South Africa) COMEC Coordenac;:ao da Regiao Metropolitana de Curitiba (Brazil) CMR Colombo Metropolitan Region CONACYT National Council of Science and Technology (Mexico) CONATA Consejo Nacional de Tasaci6n (Peru) CONDER Companhia de Desenvolvimento Urbano do Estado da Bahia (Brazil) CORETT Commission for Land Tenure Regularization CORVIDE Housing and Social Development Corporation of Medellin CPF Central Provident Fund (CPF) (Singapore) CRIC Comite de Reconstrucci6n de la Iglesia Cat6lica (Honduras) DECSAL Decentralization and Competitiveness Loan (Peru) 18 Land and Urban Policies for Porvety Reduction DFID UK Department for International Development DMC Delhi Municipal Corporation EDB Economic Development Board Singapour ESAF Escola de Administrac;:ao Fazendaria (Brazil) FAME Financial Analysis Made Easy FAT Fundo de Amparo ao Trabalhador (Brazil) FGTS Fundo de Garantia do Tempo de Servic;:o (Brazil) FIDC Fundos de Investimento em Direitos Creditorios (Brazil FIDEM Fundac;:ao de Desenvolvimento Municipal (Brazil) FJP Fundac;:ao Joao Pinheiro (Brazil) FONAVI Argentina's National Housing Fund FONCODES Fund for Community Development (Peru) FONCOMUN Trasnfer to Municipalities (Peru) FPE Fundo de Participac;:ao dos Estados (Brazil) FPM Fundo de Participac;:ao dos Munidpios (Brazil) FUNDEVI Fundacion para el Desarrollo de la Vivienda Social y Rural (Honduras) GAM Metropolitan Region of San Jose GDP Gross Domestic Product GC Gated Comunities GIS Geographic Information System GOI Government of India HBB Programa Habitar Brasil (Brazil) HDB Singapore Housing and Development Board (HDB) HDI Human Development Index HDPE High Density Polyethylene HUD CO Housing and Urban Development Corporation (India) IADB Inter-American Development Bank IBGE, Instituto Brasileiro de Geografla e Estatistica (Brazil) ICMS Impasto sabre Circulac;:ao de Mercadorias e Prestac;:ao de Servic;:os (Brazil) ILD Instituto Libertad y Democracia (Honduras) INR Indian Rupees IP Instituto de la Propiedad (Honduras). IBAMA Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renovaveis (Brazil) IBRD International Bank for Reconstruction and Development IGP fndice Geral de Prec;:os (Brazil) IMF International Monetary Fund INURBE National Institute for Social Interest Housing and Urban Reform IPEA Instituto de Pesquisa Economica Aplicada (Brazil) Land and Urban Policies for Porvety Reduction 19 IP PUC Instituto de Pesquisa e Planejamento Urbano de Curitiba (Brazil) IPTU Impasto Predial Territorial Urbano (Brazil) IPV Provincial Housing Institutes (Argentina) ISGM Informal Settlement Growth Model LDC Less Developing Country LUD Law of Urban Development (Mexico) MCD Municipal Corporation of Delhi MCR Mega City Region MDF Municipal Development Fund MIC Middle Income Country MAG Metropolitan area of Guadalajara MRBA Metropolitan Region of Buenos Aires , MUD PA Ministry of Urban Development and Poverty Alleviation NCAER Council of Applied Economic Research NIE New Institutional Economics NGO Organizac;:ao Nao Governamental (Brazil) NSDP National Slum Dwellers Program (India) OECD Organizac;:ao para Cooperac;:ao e o Desenvolvimento Economico (Brazil) OODC Outorga Onerosa do Direito de Construir PDG Population Density Gradient PETT Registration of Rural Property Program (Peru) PNAD Pesquisa Nacional par Amostra de Domidlios (Brazil) PPP Parceria Publico-Privada (Brazil) PREZEIS Plano de Regularizac;:fo das Zonas Especiais de lnteresse Social (Brazil) PRIMED Program of Regularization of Low-income Settlements PROAP Low-income Settlements Urbanization Program (Brazil) PROCEDE Ejidal Rights Certification Programme (Mexico) PRONAA Programa Nacional de Assistencia Alimentaria (Peru) PROSANEAR Programa de Saneamento para Populac;:oes em Areas de Baixa Renda (Brazil) PRO VIAS Program of Road Construction (Peru) PTO Permits to Occupy (South Africa) RIDE Regiao Integrada de Desenvolvimento (Brazil) RMC Regiao Metropolitana de Curitiba (Brazil) RMR Regiao Metropolitana do Recife (Brazil) RUS Regimen Unico Simplificado (Peru) SABESP Companhia de Saneamento Basico do Estado de Sao Paulo (Brazil) SANAA Servicio Aut6nomo Nacional de Agua y Alcantarillado (Honduras) SATs Servicio de Administraci6n Tributaria (Peru) 20 Land and Urban Policies for Porvety Reduction SDUV Sub-Secretarfa de Desarrollo Urbano y Vivienda (Mexico) SENA National Skill Training Institute (Colombia) SFH Sistema Financeiro de Habita<;:ao (Brazil) SIDA Swedish International Development Assistance SIPP Survey of Income Program and Participation (USA) SME Small and Medium-Sized Enterprises STN Secretaria do Tesauro Nacional (Brazil) SUNARP National Register Entity (Peru) TDR Transfer Development Rights UBN Unsatisfied Basic Needs of the Population ULA Urban land Act (Honduras) ULO Urban Land Organisation (Honduras) UN United Nations UN CHS United Nations Centre for Human Settlements - Habitat UNFPA United Nations Population Fund UIT Unidad Impositiva Tributaria URIF Urban Reform Incentive Fund (India) VAMBAY Valmiki Ambedkar Awaaz Yojana (India) WHO World Health Organization ZACs Zones d' Amenagement Concerte (France) ZEIS Zonas Especiais de Interesse Social (Brazil) PREFACE LAND, SHELTER, TRANSPORT: THE LATIN AMERICAN WAY Peter Hall We're just passing one of the great milestones in human history - but hardly anyone is noticing. It isn't anything outwardly dramatic, like a revolution or a war. But it is fundamental, in the sense that the Industrial Revolution in Britain was fundamental. Future historians, doubtless, will call it the Urban Revolution. For the first time in history, a majority of the world's six billion people are living in cities. Between 2000 and 2025, on the best estimates we have from the United Nations, the world's urban population will double, to reach five billion; city-dwellers will rise from 47 per cent to over 61 per cent of the world's population. But that's not all. Most of this explosive growth will occur in the cities of the developing world. There will be a doubling of the urban population, in the coming quarter century, in Latin America and the Caribbean, in Asia and in Africa together - above all in Africa. Even by 2015, the UN predict that there will be 3 5 8 "million cities'', with one million or more people; no less than 153 will be in Asia. And there will be 27 "mega-cities'', with ten million or more - 18 of them in Asia. It is here, in the exploding cities of some of the poorest countries of the world, that the central challenge lies. - Table 1: World Urban Population, 1980-2000-2020 Urban Population Urban Population in % Urban Population Growth Rate in % ~-""""--.~~~--' 1980 "' 2000 2000 1980-85 2000-05 2020-25 World 39 47 57 2.6 2.2 1.7 Africa 27 38 49 4.4 4.0 3.0 Europe 69 75 80 0.8 0.3 0.1 North America 74 77 82 1.2 LO 0.9 Central America 80 67 73 3.1 2.0 1.5 South America 68 80 85 3.1 1.8 1.1 Asia 27 38 50 3.6 2.0 2.0 Oceania 71 70 72 1.4 1.3 1.3 Developing Countries 29 41 52 3.8 2.9 2.1 Developed Countries 71 76 81 0.9 0.5 0.3 Source: World Resources 1998-99 22 Land and Urban Policies for Porvety Reduction A huge challenge, to be sure - but also a huge range of opportunities: opportunities for greater freedom, greater freedom above all for development, as people leave behind their traditional bondage to the land and the total dominance of the daily struggle for food. Urbanization is a fundamental form of liberation of the human spirit: in the famous German quotation from the Middle Ages, Stadt!uft macht Frei: the city air makes you free. It does more than that: just because it frees up human creativity, the city is the place where the great advances occur - artistic, intellectual, technological and also organisational. You need urbanisation if you're going to get development. City growth is potentially a great thing. But only potentially. Urbanization is a basic precondition for development. But it doesn't of itself guarantee development. There's good urban growth and there's bad urban growth. Managing urban growth so that it contributes positively to economic advance, reconciling it with ecologically sustainable forms of development and reducing social exclusion, represents the key challenge for urban planners and urban managers in this new century. The Fundamental Challenge The major challenge, for all those of_us who care about cities, comes from the burgeoning ci~ies of the developing world, where there is a paradox: people are still flooding into these cities, too many children are being born in those cities based on the hope for a better life; but too often they are being cheated. For urban growth has brought a sharp rise in urban poverty: according to UNFPA estimates, over one in four of the people in the cities of the developing world lives below official poverty lines, and that proportion rises to more than one in three in the Middle East and North Africa and to more than two in four in sub-Saharan Africa. And a large proportion of the poorest are women. In these cities, the quality of the environment is not improving; in far too many cases, it is deteriorating. The problem is daunting. Many of these cities are already bigger than their equivalents in the developed world, and are projected to become yet larger. Most have only recently started on their development process. And, with some conspicuous exceptions, they lack the governmental structures and the administrative traditions to tackle the resulting problems. Let's be fair: they have achieved a great deal against overwhelming odds; and some have emerged as models for the rest of the world. But they are too few, and their example is not spreading fast enough. Three Kinds of City: Three Kinds of Problems However, and this is the first important point I want to make this morning, the term "developing city", like the term "developing country", is no longer very meaningful. In fact, I want to argue that it's fundamentally confusing. The World Commission on 21st-Century Urbanism, which presented its report Urban Future 21 to a major conference in Berlin in the year 2000 (Hall and Pfeiffer 2000), argued that we can most usefully divide cities worldwide into three major categories, and that so-called "developing cities" in fact fall into two different categories. Even this is crude and simplistic, but it makes the point. The first the Commission called the City coping with Informal Hypergrowth. It is represented by many cities in sub-Saharan Africa and in the Indian subcontinent, by the Moslem Middle East, and by some of the poorer cities of Latin America and the Caribbean. It is characterized by rapid population growth, both through migration and natural increase; an economy heavily dependent on the informal sector; very widespread Land and Urban Policies for Porvety Reduction 23 poverty, with widespread informal housing areas; basic problems of the environment and of public health; and difficult issues of governance. The second they called the City Coping with Dynamic Growth. It is the characteristic city of the middle- income rapidly-developing world, represented by much of East Asia (including China), some of South Asia, much of Latin America and the Caribbean and the Middle East. Here, population growth is falling, and some of these cities face the prospect of an ageing population. Economic growth continues rapidly, but with new challenges from other countries. Prosperity brings environmental problems. The City coping with Informal Hypergrowth In this first kind of city, the key problem is that the urban economy can't keep pace with the growth of the people. There are high birth rates - a product of sexual ignorance, superstition and above all poorly-educated, often illiterate women. This, plus continued migration from the countryside, produces a huge surplus of unskilled labour. Many of the migrants have been pushed off the land rather than positively pulled into the cities, by famine or civil war or insurrection: too often, they are virtually starving. They go into the only work they can find, in the informal economy: casual work and petty trading. This leaves them in dire poverty - especially the women and above all the female-headed households, which typically form more than 30 per cent of the poor population. The problem is that in these cities the formal or modern sector is too often struggling to survive, and too often giving up the battle. This is particularly true of indigenous enterprises. They can't compete, for multiple reasons: under-education, poor infrastructure, lack of credit and failure to access global markets. So you find cities that - apart from global enterprises like hotel chains or fast food outlets - lack a formal economic base, cities in which the great majority of people live in informal slums, often in very bad conditions, and eke out an existence in the informal economy. They have little work and they live at the margin of existence, in places that lack the basics for a civilised life. They have no respect for the environment, because they can't afford to do anything except struggle for survival: if keeping warm means cutting down the remaining trees for firewood, they'll do it; if keeping alive means drinking polluted water, they'll do that. And they find it hard to contact to worthwhile jobs, even if they had the skills, because they can't physically reach them: lacking either a bicycle or a bus fare, they have nothing but their own two feet. If you visit such cities, your first reaction may well be despair. But there is actually a solution to this huge raft of problems, though it may sound paradoxical. First, it's to get the birth rate down, which means basic education, above all education for the girls. Our report argues that there's a tremendous role for information technology here, if we can get low-cost machines that don't need to depend on erratic mains electricity. In fact technology has taken a huge leap even in the five years since we were working on our report, through the development of battery-powered mo bile phones that can hook up directly to the Internet. And this is just the beginning. Then, the key is progressively to formalise the informal economy. Cities can do this in various ways: strengthening relationships to the mainstream economy, both for inputs and outputs - for instance, through schemes to provide microcredit, providing building materials and food and water, and more effective transportation to help people access a wider range of jobs. They can achieve this best through communal self- 24 Land and Urban Policies for Porvety Reduction help neighbourhood projects, backed up by informal levies to pay for materials, which can help overcome bottlenecks in basic infrastructure. Microcredit schemes, providing tiny loans so people can start their own businesses, will play a particularly crucial role. The City Coping with Dynamic Growth This kind of city is important for this conference, because most Latin American cities belong to it. Here there's good news: the trend is for population growth to fall sharply, because of urbanisation, as people see that the costs of education and rearing children rise while the economic value of children goes down. (These are two sides of the same coin: crudely, the value of uneducated young people tends to decline, so it simply takes much longer and costs more to get them to the point when they become effective earners). And this has a further knock-on effect: there is a big rise in the number of working-age people relative to the young and the old, who have to be looked after. In the jargon, the dependency ratio falls to a minimum. So that's the good news, and it isn't the end. In these cities, the great passage from the informal to the formal economy is already well under way. Many of them are very attractive to inward investment, because they offer a well-educated and well-trained labour force at lower wages than in developed cities, and besides economic growth is generating big domestic markets for consumer durables like cars and refrigerators and personal computers. China· is the outstanding case here, following on a hugely bigger scale the example earlier set by "tiger economies" like Singapore, Hong Kong or South Korea. But there's a sting in the tale there: this foreign direct investment can always be diverted to even lower-cost countries and cities, as some Latin American cities are now finding to their cost. The key is to keep trading up into more sophisticated levels of production, especially advanced services, as both Singapore and Hong Kong have done during their forty years of sustained growth, and as leading Chinese cities like Shanghai are now doing. The main result of all this is that cities in this group all find themselves in a state of quite extraordinary dynamism but also of rapid transition. It often seems as if they're going through every stage of economic development at once. Or rather, different sections of their population are going through different stages. Side by side, in the downtown business districts you can see gleaming new high-rise office towers housing global corporations that provide advanced business services; along the arterial expressways, sleek suburban factories that are pouring out consumer goods as well as forests of new apartment towers; and, in between, wretched informal slum settlements where the people struggle to make a basic living by performing odd jobs or selling trinkets. These cities often look as if they're simultaneously first world cities and third world cities. One result is that they are highly polarised. Many of them, though not all, display extraordinary contrasts in wealth and poverty. Cities in South Africa and Brazil, two of the most unequal countries on earth, display this pattern to an extreme degree - but it's now observable in China and in Poland. A significant sign is to see heavily gated, even armed luxury apartment blocks or country-dub type developments, next to wretched shacks or worn-out slum apartments. All too often, in many though not all of these cities, there are reports of escalating crime and violence. The poor, some of them, may find solace in drink or drugs, compounding the problem. Because the poor have to find somewhere to live, they often contribute to environmental disasters by building their homes on unstable hillsides or on floodplains, with results that are sometimes tragic. Even Land and Urban Policies for Porvety Reduction 25 when they and their homes survive, they are often located far from job opportunities, with poor or non- existent bus services, compounded by traffic congestion. The answer to these problems is to continue to push the economy in the direction first of advanced manufacturing and then of advanced services, always keeping one step ahead of the global competition. (Again, Eastern Asian cities provide the classic model). Of course, cities cannot provide all the necessary policies on their own: nation states have to provide the right framework of macro-economic policies. But cities can do a lot, especially if they are given the right degree of administrative and fiscal autonomy- which many of them have been getting, already, during the last twenty years. Above all, they must and they can help their poorest citizens to join the mainstream economy and the mainstream society. Then and Now ... It's helpful at this point, I think, to turn from a geographical kind of comparison to an historical-geographical comparison. In some important ways, not least I come levels, cities in this group compare with cities in the mature developed world about a hundred years ago. London, Paris, Berlin, New York in 1905 can be compared with Sao Paulo, Mexico City, Caracas and Bogota today. Both groups of cities were, or are, growing explosively both in population and wealth. Both displayed, or display, extreme divisions of wealth. Both contained, or contained, huge high-income areas of great affluence and also huge slum areas of great wretchedness. But there are, I would argue, two key differences. The first is in housing. Then, the slums had a formal characteristic: they were of permanent construction, generally large houses built for wealthy people (as in London), sometimes apartment blocks (as in Paris or New York), subdivided and sometimes again subdivided, and therefore chronically overcrowded. Now the corresponding slums are informal: self-built and unserviced. In fact, they correspond very precisely to the slums of the first category of cities, which shows us that this second category is really an amalgam of the first type and the fully-developed mature city. The second key difference was, or is, in transport. The basic reason for the slums of 1905 was that the poor, who depended on informal employment, had to crowd ever more closely into housing near their work - that is, in or near the city centre. In London at that very time, the great social reformer Charles Booth wrote a paper ·entitled Improved Means of Locomotion as a first Step towards the Cure of the Housing Difficulties ofLondon (Booth 1901). And in fact, just that was happening. Already, London had the world's first underground railway; in 1900, it was already nearly forty years old. And, aided by American capital, the tunnelling teams were burrowing under London's streets. Most of the tube network, on which you travel if you visit London today, was built by the year 1907. And simultaneously, the municipal authority for London, the London County Council, was electrifying and extending the tramcar system to serve new public housing estates, offering very low workmen's fares so that poor people could afford to live in good housing on the edge of the city while getting to their jobs in the centre. Latin American cities today, in contrast, are in some cases very much larger - the Sao Paulo metropolitan area is three times the size of London one hundred years ago - yet have much less well- developed public transport systems. The paradoxical, even perverse, result is that relatively speaking, the poor in these cities have much greater problems in getting to work than their counterparts in London or New York in 1905. 26 Land and Urban Policies for Porvety Reduction Housing in the Developing World How adequate is housing in the developing world? UN-Habitat figures show a mixed picture. Very evident is the fact that two areas - Latin America and the Caribbean, and Asia - show far better standards than Sub- Saharan Africa or North Africa and the Middle East. The same is evident for provision of basic infrastructure like water, sewerage, electricity or telephone service. To a remarkable degree, throughout the developing world, most housing is well-serviced. But for informal housing, the position varies very much. Generally, however, provision in Sub-Saharan Africa falls well behind that in the rest of the d~veloping world. That raises the basic question; what is sub-standard housing? How do we define a slum? UN Habitat has sought to produce a rigorous, generally-applicable definition. They use five key elements: access to water, access to sanitation, structural quality of housing, overcrowding and security of tenure. Using that as the basis, Table 2 from UN-Habitat shows the relative proportion of slum housing by region, worldwide, in 2001. Overall, slum dwellers constitute 32 per cent of the world's urban population. For developing countries, the figure is 43%; for the least developed countries, 78 percent. This represents a huge differential between Sub-Saharan Africa and the rest of the developing world. Table 2: Distribution of the World's Urban Slum Dwellers, 2001 Region Urban % in total % slum dwellers population population in total urban (000) population Sub- Saharan Africa 231,052 34.6 71.9 Asia Pacific 1,211,540 35.4 43.2 Latin America and Caribbean 399,385 75.8 31.9 Middle East and Northern Africa 145,624 57.7 29.5 Transition economies 259,091 62.9 9.6 Advanced economies 676,492 78.9 5.8 World 2,923, 184 47.7 31.6 Developing countries 2,027,665 40.9 43.0 Least developed countries 779,239 26.2 78.2 Source: UN-Habitat, 2003a. Slum development is systematically associated statistically with GDP per capita and with the UN's Human Development Index. But there is a striking systematic relationship between the prevalence of slum housing and inequality of income (rather than absolute income), as Table 3 shows. The UN-Habitat analysis suggests that generally throughout the developing world, despite rising per capita income levels, housing is becoming less rather than more affordable, both for owners and renters. But there are major differences between the least and the most developed regions: Latin America appears quite highly developed in terms of housing affordability, suggesting that the process of formalising informal settlements has been successful overall. Rather remarkably, most inhabitants of informal housing do not squat rent-free, but pay rent to a landlord. This suggests the degree to which there is an incentive to own. Land and Urban Policies for Porvety Reduction 27 Table 3: Slums and Income Inequality Country Income ratio Slum dwellers (richest 20% to (%of urban poorest 20%) population) Sierra Leone 57.6 96 Nicaraqgua 48.8 81 Guatemala 46.0 62 South Africa 45.9 33 Lesotho 43.4 57 Honduras 42.7 18 Nigeriaq 40.8 79 Cameroon 36.6 67 Kenya 36.1 71 Cambodia 33.8 72 India 33.5 55 Central African Republic 32.7 92 Bolivia 32.0 61 Morocco 30.9 33 Lao People Democratic Republic 30.6 66 Ghana 30.1 70 .Sauce: UN-Habitat, 2003b; UNDP, 2002. Housing and Transport: The Pacific Asian and Latin American Ways One important key for the people in such areas is to help them formalise their housing: to use communal self-help to provide the necessary infrastructure, so that they begin to turn their informally-built areas into middle-class neighbourhoods. In countless Latin American cities, it has been happening and is still happening. In many eastern Asian cities, the approach has been different: the city itself has intervened to tear down informal neighbourhoods and provide high-quality housing, first for rent, later for sale, either through public provision or, increasingly, by policies that foster the growth of owner-occupation, as now in Singapore. There is no one right way here; there are different paths towards the same goal. The UN-Habitat 2003 report contains a number of urban case studies, several located in Latin America. Bogota demonstrates forty years of "informal" growth - here, mainly not due to squatting, but to illegal subdivision. Vast settlements such as Ciudad Bolivar, Bosa and Usme at first lacked water, drainage, sewerage, power, education and health care. But they saw consistent improvement, in which the city authorities worked collaboratively with local inhabitants (UN-Habitat 2004, 88). Here in Bogota, which is characterised by a special form of low-income neighbourhood called the barrio pirata (pirate neighbourhood), formed not through land invasions through an informal process of land subdivision and granting of title, there has recently been a huge "de-marginalisation mega-project", which between 1998 and 2000 used a budget of US$800 million to construct 110 kilometres oflocal roads, 2300 kilometres of drainage, six hospitals, 51 schools, 50 parks, four major public libraries and legalising 450 28 Land and Urban Policies for Porvety Reduction settlements. It did not fully achieve these targets, falling significantly short on surfacing and lighting of roads, partly because it depended on the sale of a telephone company that failed to go through- but it is nevertheless impressive. The problem, as in so many other Latin American cities, is that though the city achieved measurable and significant improvements on key measures 1, none the less poverty rose sharply (from 35% below the official poverty line in 1997, to 49.6% in 2000) and income inequalities grew as more and more internal refugees flood into the city escaping political violence outside, causing new household formation to surge ahead of housing provision (Skinner 2004, 80-1). The Sao Paulo case study demonstrates that here, there are two distinct kinds of slum: corticos (rented rooms in subdivided inner-city tenements), of very poor quality but close to jobs and urban services, and f"avelas, found everywhere, but for the fact that in the city itself, private owners tended to regain possession of squatted areas - two only survive here, both very large (Heliopolis and Parais6polis) but the great majority are now found in the poorest, peripheral, environmentally-fragile areas (UN-Habitat 2004, 89). Mexico City produces two case studies in the UN-Habitat 2003 report. The first, Nezahualc6yotl concerns a huge irregular settlement that developed from the 1950s on a drained lake bed outside the federal District. Here, legal title was ambiguously legal: speculators "sold" plots and the state government subsequently regularised title. But the resultant developments lacked basic services such as paved roads, lighting, water and main sewerage. From the end of the 1960s a citizens' movement, Movimiento Restuarador de Colonos, successfully campaigned to secure progressive legalisation of titles and basic servicing, even extending, at the Millennium, to extension of the Metro outside the Federal District. As a result, by the end of the 1990s, only 12% of the area was still held in irregular title. But the quality of basic services varies greatly: 63% of households have inside water supply, but 15% still have poor roofing (UN-Habitat 2004, 94). The second Mexico City case study concerns the Valle de Chalco Solidaridad, a vast informal settlement south east of the federal District. This was an agricultural area, where in the early 20th Century, after the Mexican revolution, the land was expropriated and given to the peasants. But after 1950 the plots became uneconomic to farm at just the time when, resulting from urban sprawl, the land became attractive to speculators. The land was subdivided and sold on credit, and between 1970 and 2000 the population rose from 44,000 to 323,000. Here, too, by 1998 90% of the plots had regularised title, and major infrastructure had taken place. Even so, at that date basic housing conditions remained very bad: 78% of households had no inside tap, 40% still had cardboard roofing, 20% lived in one room (UN-Habitat 2004, 91). The conclusions from these UN-Habitat case studies are very clear, and they give mixed signals. Informal settlement tend quite rapidly to become regularized, and their inhabitants to receive legal title, while services are progressively provided: first basic ones like piped water, sewers, paved streets and street lighting, then more advanced services like schools, libraries and even Metro service. But the resultant provision is still incomplete, with different standards. Meanwhile, the entire invasion/improvement process ripples ever farther out from the urban core, bringing a problem of access to jobs, with long commuting distances and even longer times. As a result, the quality of transport service becomes crucial. 1 Between 1993 and 2001, the percentage lacking more than one of five key measures - inadequate housing, lack of drinking water or sewerage, overcrowded shelter, non-attendance at school by at least one child in the household, and dependence of more than three household members on a head with less than four years of primary school - fell from 3.5 to 2.4. Land and Urban Policies for Porvety Reduction 29 Here, too, there is a basic difference in approach. Some Eastern Asian cities have deliberately encouraged high-density development which will support a top-quality metro system - and some, like Hong Kong and Singapore, had no choice because they had so little fand. China seems to be going the same way, as can be seen in Shanghai. Some Latin American cities, in contrast, have made extraordinary innovations in operating bus systems to serve their more far-flung residential neighbourhoods - and one of the most extraordinary of all, Curitiba in Brazil, has created a bus system that works like a metro, with local buses that feed into an express system travelling on its own tracks; Bogota in Colombia has developed a very similar system. Latin American cities, above all Brazilian cities, have taken a world lead over the past thirty years in developing highly innovative urban bus-based transit systems. For this there have been very good reasons. As we have seen, rail-based metro systems have b~en far less developed, especially 30 years ago; Brazilian cities simply lacked the resources for expensive tunnelled rail systems, and made a virtue out of necessity. Curitiba's "Bus Metro" system was the great pioneer, widely hailed and now widely imitated in cities as diverse as Bogota, Sao Paulo and many others. Brazilian engineers took the lead in developing these solutions. But at their best they involved not just engineering but also planning approaches, since they integrated bus service and land use planning. The central feature of the Curitiba system is a variety of services - express buses running along special bus corridors, orbital services and local services, all integrated through high-speed transfer stations at a variety of points all over the city, and used as the basis of a land-use policy that encourages high-density development and redevelopment along the express corridors. The buses on the express corridors are very high-capacity bi- articulated vehicles with a total capacity of 270, more akin to a light rail train than an ordinary bus. Painted red, they interchange at the transfer stations with buses running on orbital routes from suburb to suburb, painted green, and with local feeder or "conventional" buses painted yellow. The comparative capacities of the buses on the different systems vary greatly. All are operated privately on a franchised system. The express corridors have been deliberately .developed through planning and zoning controls for very high-density, high-rise mixed development - as is very .evident from the tourist's view from the top of the city's television tower. Thus Curitiba's success became a Brazilian success. Brazilians make over 60 million bus trips a day; Americans, living in a country with twice the urban population, make only one third as many. Brazilian cities demonstrate some of the highest rates of bus ridership in the world: Sao Paulo and Rio between them have about as many daily bus journeys· as the entire United States, which has ten times their combined population. All the major Brazilian cities have made major innovations in bus operation: in the 1970s, Sao Paulo and Porto Alegre pioneered the idea of running buses in convoys along a dedicated lane, and Porto Alegre developed an integrated paratransit system. These innovations were driven by necessity: bus-based transit systems average $5 million per mile ($3 million per kilometre).against $20-$100 per mile ($12-62 per kilometre) for light rail or metro systems. The success of these bus-based solutions - urban bus operations in Brazil yield positive net revenues of over $3. billion per year - have created a flourishing export industry, with worldwide consulting operations; the engineer Pedro Szasz, developed the bus convoy systems in Sao Paulo and Porto Alegre, engineered the combination oflocal, skip-stop and express services that constitute the Transmilenio Bus Rapid Transit (BRT) in Bogota (Golub 2004, 4-5; Skinner 2004, 78). But there's a funny point: if you visit Singapore and Curitiba, the two cities look very alike, because both have integrated their land use and transportation policies, encouraging high-density and high-rise development along their main trahsportation corridors. Again, there's mpre than one way towards the same goal, but in the end the outcom~s may be very similar. 30 Land and Urban Policies for Porvety Reduction It's no accident, perhaps, that Curitiba and Singapore are now two of the richest cities in this group; in effect both have made the transition into the developed world, and both are technologically and organisationally among the world's most advanced cities. These cities are leading their countries in technological and organisational innovation, showing the way for other cities either to imitate them or to go in a different, equally innovative, direction. That is the path of rapid development. There are some important conclusions, therefore, regarding transport. Latin American cities demonstrate that bus-based cities do work: they can deliver good service, with high passenger volumes, at remarkably low cost. ·But there is a basic question. Can they do so everywhere - especially, to the urban periphery? If they fail to do this, is the urban transport problem in the largest cities destined to become steadily worse? I want to argue that it will not, because of the emergence of a new urban phenomenon: the Mega-City-Region. A New Urban Phenomenon: The Mega-City-Region Another key difference between the great cities of a century ago, and now, is this new phenomenon: the Global Mega-City-Region. This is a pattern of extremely long-distance deconcentration stretching up to 150 kilometres from the centre, with local concentrations of employment surrounded by overlapping commuter fields, and served mainly by the private car. The Pearl and Yangtze River Deltas in China and South East England, around London, are two of the world's leading examples of this phenomenon. In Pacific Asia, it has recently been predicted that by 2020 two-thirds of the population of the ASEAN group of countries will be found in only five MCRs: Bangkok (30 million), Kuala Lumpur-Klang (6 million), the so-called Singapore Triangle (10 million), Java (100 million) and Manila (30 million). In adjacent Eastern Asia, these agglomerations are even bigger: Japan's so-called Tokaido corridor (Tokyo-Nagoya-Kyoto-Osaka-Kobe) is predicted as having a total population of 60 million, China's Pearl River Delta (Hong Kong-Shenzhen-Guangzhou) 120 million, and the Yangtze River Delta (Shanghai-Suzhou-Hangzhou-Nanjing) 83 million (McGee 1995, Wo-Lap 2002, quoted in UN-Habitat 2004, 63). The precise spatial details vary from country to country according to culture and planning regime, and for this reason population figures and predictions should be treated with caution, but the pattern is emerging very clearly and very rapidly around some of the largest cities in this second category: it is very evident around Sao Paulo, and has recently been analysed in some detail by Adrian G. Aguilar and Peter M. Ward for Mexico City (Aguilar and Ward 2003). Latin America is highly urbanised. In 2000, in Latin America and the Caribbean, 75 .4% of the total population, 400 million, were urban; 31.6% of the total population, 41.8% of the urban population lived in cities of more than one million, while 15.1 % of the total, 31.5% of the urban population, lived in metros with 5 million and more people. And these included some of the biggest urban agglomerations in the world: Mexico City, with 18.1 million, 2nd; Sao Paulo, with 17.9 million, 3rd; Buenos Aires, with 12 million, 11th; and Rio de Janeiro, with 7.4 million, 15th. Also in this list were Bogota (6.8 million) and Santiago (5.5 million) (UN- Habitat 2004, 64). However, it is extremely important that the term "city", in this sense, is not the administrative entity but a much larger metropolitan area. In the largest cases, such as Mexico City and Sao Paulo, it is in fact an equivalent of the Asian mega-city region. These mega-city-regions develop through a complex process of Land and Urban Policies for Porvety Reduction 31 simultaneous decentralisation at a regional scale, and recentralisation at a more local scale: a process that Dutch planners in the 1960s called "concentrated deconcentration". Thus they are increasingly polycentric. In recent decades, it has been observed that central city growth has slowed while peripheral growth has speeded up. As the UN-Habitat 2004/5 report notes, " ... significant shifts from city-centred to regional forms of urbanization are currently taking place" (UN-Habitat 2004, 65): multi-nodal, urban regional systems are developing, in which new sub-centres are independent in terms of their social and economic patterns, but are functionally linked to the big city, a process that in a recent European study we have termed functional polycentricity (Hall and Pain 2004). In the Mexico City metro, more than half the population lives outside the central Distrito Federal which is generally regarded as the city. In Sao Paulo, the city contains 10 million people, just half that found in the wide metropolitan area (19.8 million). In Buenos Aires, out of a total metropolitan population of 12 million, only 3.5m live in the Capital Federal (UN-Habitat 2004, 65-66). Failure to appreciate or understand this process has led to some quite serious errors. In the 1970s, urban analysts incorrectly predicted further explosive growth of metro areas: Mexico City for instance was predicted in UN publications as growing by the year 2000 to 30 million. In fact, almost as these predictions were being made, growth tapered ~harply and stopped at the 20 million point. There were two reasons for this, neither having conspicuously much to do with planning. First, because of obvious emerging negative externalities in l the Mexico City metro, migrants from rural areas diverted to second-order cities such as Guadalajara and Monterey. Secondly and even more significantly, within the general ambit of Mexico City growth diverted to "secondary cities" at increasing distances, many informal settlements of vast size such as Nezahualc6yotl and Ecatapec, located in the adjacent State of Mexico (UN-Habitat 2004, 50, 65). Aguilar and Ward show that Mexico City's Federal District is now merely the core of a huge and polycentric mega-city-region stretching up to 100 kilometres and more from the Z6calo. In fact more than half the population of the region is now found outside the District. Over the last 35 years, population growth has rippled out in concentric circles at steadily increasing distances from the city centre, and the most rapid growth is now in the peripheral areas. This outer zone is characterised by huge informal settlements like Ecatapec and Nezahualcoyotl, with up to one or two million people apiece. Very significantly, these settlements suffered from serious deficiencies in basic infrastructure thirty years ago, but had largely caught up by the 1990s (Aguilar and Ward 2003, passim). I will return to that point a little later. Equally important however is another point: these outer areas are not just vast residential zones. They now contain economic subcentres which are increasingly important in their own right. And in this process, which could be called the increasing polycentralisation of the region, there is an increasing specialisation of function: the more advanced or formal parts of the economy remain within the Federal District, even in its core, while the outer centres attract manufacturing and retail functions. To the north these are dominated by heavy, large-scale and high-tec~nology enterprises such as metallic and chemical industries; to the east, they are dominated by small-scale informal activities; in some parts of this zone, significantly, there was a decline in employment in traditional craft industrial employment. But there was also a notable growth of tertiary activity in this zone along major transportation corridors (Alguilar and Ward 2003, 15-16). The process has distinct advantages. As jobs develop in the outer rings of these metropolitan areas, the burden of commuting can lessen. In Bogota, though population grew 40%, it was found that travel distances stayed the same (UN-Habitat 2004, 52). 32 Land and Urban Policies for Porvety Reduction The Basic Emerging Problem: Governance in the Mega-City-Region There is currently a basic problem with all these Mega-City-Regions: they suffer from fragmented governance. The Mexico City metropolis has 28 municipalities, and more than half the population lives outside the Distrito Federal. The Sao Paulo metro is similarly divided among 39 districts/municipalities; Rio de Janeiro among 13 municipalities, and Buenos Aires among 20 municipalities that enjoy varying degrees of autonomy; the Curitiba metropolitan area is governed by no less than 25 municipalities (UN-Habitat 2004, 58, 66). This last case is particularly significant: within the Curitiba metropolitan area the population of the city accounts for only 61 % of the population - and is falling. And, despite the legendary worldwide reputation of the city for delivery of highly innovative services, the evidence from the wider region is far less encouraging: 500,000 live below the Brazilian official poverty line, there are 89,000 substandard units in 903 problem housing areas, only 58% of the area is sewered and only 35% of the sewerage is treated. A regional planning authority, COMEC, has existed for nearly twenty years and has generated plans but no action, because it has no effective powers (Macebo 2004, 547-8). In conclusion, therefore, the overwhelming need in all these great metropolitan areas is for effective metropolitan governance across the entire mega-city-region. Such regions are the new reality of urban existence in the 21st century. They are, as earlier said, both the solution and the emerging problem. They are a Solution because the offer the prospect of re-equilibrating homes, jobs and transport across a new and vast spatial scale. But they are also the Problem because this demands effective planning, powers and action across a very wide Metropolitan scale. Unless this opportunity can be grasped, the evident risk is that such regions will be characterised by a deepening economic and social imbalance and polarisation, bet:Ween rich central cities and marginalised poor peripheries. The signs are already evident. There is some time to grasp the problem and resolve it - but, perhaps, less than we think. References Aguilar, A.G., Ward, P.M. (2003) Globalization, Regional Development, and Mega-City Expansion in Latin America: Analyzing Mexico City's Peri-Urban Hinterland. Cities, 20, 3-21. Booth, C. (1901) Improved Means ofLocomotion as a first Step towards the Cure of the Housing Difficulties of London. London: Macmillan. Golub, A. (2004) Brazil's Buses: Simply Successful. Access, 24, 2-9. Hall, P., Pfeiffer, U. (2000) Urban Future 21: A Global Agenda for Twenty-First Century Cities. London: Span. Macebo, J. (2004) City Profile: Curitiba. Cities, 21, 537-550. McGee, T.G. (1995) Metrofitting the emerging Mega-Urban Regions of ASEAN. In: McGee, T.G., Robinson, I (ed.) The Mega-Urban Regions ofSoutheast Asia. Vancouver: University of British Columbia Press. Skinner, R. (2004) City Profile: Bogota. Cities, 21, 73-81. UN-Habitat (2001) Cities in a Globalizing World: Global Report on Human Settlements 2001. London and Sterling, VA: Earthscan. Land and Urban Policies for Porvety Reduction 33 UN-Habitat (2003) The Challenge ofSlums: Global Report on Human Settlements 2003. London and Sterling, VA: Earthscan. UN-Habitat (2004) The State ofthe World's Cities 200412005: Globalization and Urban Culture. London and Sterling, VA: Earthscan. Wo-Lap, Lam, L. (2002) Race to become China's Economic Powerhouse. CNN, 11June2002. GENERAL INTRODUCTION INTERNATIONAL URBAN RESEARCH SYMPOSIUM 2005 Mila Freire, Bruce W Ferguson, Ricardo Lima, Dean Cira and Christine Kessides As Sir Peter Hall notes in his preface, the "Urban Revolution'' now occurring largely in developing countries presents great opportunities and risks. Urbanization can help raise standards of living, provide the infrastructure and services for immense improvement in human welfare, and free people from the bondage to land and total dominance of the daily struggle for food. The attractive n~ighborhoods and downtowns, efficient transport, many amenities, impressive social indicators, and high standard ofliving of Singapore and Curitiba signal this potential. However, if mismanaged, the urban wave can bring a sharp rise in urban poverty, result in surrealistically desperate conditions, and foment disease and violence. The pavement dwellers of Mumbai living cheek by jowl with the immense wealth of this commercial capitol of a newly-prosperous India, and the seemingly endless slums and hovels that consume many sub-Saharan African cities are emblematic of this other urban present and possible future . Urban land lies at the center of many of these opportunities and risks. Assembling reasonably priced, well-located . land parcels has become the most crucial challenge for affordable housing development. Wh~n - as is often the case - such programs are unavailable, the low/moderate-income majority in many developing country cities usually cannot afford to purchase the least-expensive commercially-built home and, instead, use "informal" systems to house themselves. Such "progressive housing" also starts with and depends on access to a lot. Similarly, efficient transport and the ability of households to connect with jobs and services depend on land-use and density. Near the start of the great urban wave in developing countries - in the 1950s - poor households migrating to cities from the countryside could, with some frequency, find centrally-located low-cost land on which to settle. The film "Black Orpheus" that re-creates the myth of Orpheus and Eurydice in the shantytowns on the steep hills with panoramic views above Rio de Janeiro paints an idyllic picture of Favela life at this time. It is impossible to imagine that such a lyrical film on favela life would be made now. Indeed, the Brazilian cinema currently produces many gritty, neo-realistic films featuring the blow-back from the spread and worsening conditions in favelas, including street orphans, kidnapping, and urban violence. In this regard, the era of easy access to urban land is long gone in most developing-country cities. Continuing urbanization has used up the most developable areas around many cities. Although government agencies frequently own some land in urban and peri-urban areas, large development companies that build mainly for middle and upper-income households now appear to own most of the remaining developable parcels. The low rates, high technical requirements, and political difficulties of the real property tax in developing countries allow such large landowners to continue to hold their parcels at little cost. Without mitigating measures, land titling and other market reforms have resulted in the "commodification" of land and housing (Durand- Laserve), often raising prices and excluding the poor. For many reasons, urban land has now become the main "binding constraint" to housing the poor. 36 Land and Urban Policies for Porvety Reduction This anthology collects and organizes 32 papers presented at the International Urban Research Symposium held on April 4-6, 2005 in Brasilia focused on urban land. IPEA and the World Bank jointly sponsored this event. The papers presented at this 2005 Symposium have been organized around six key themes in the sections of this anthology: • Land markets, land development, and land policy • Secure tenure, property rights, and informal land delivery systems • Informal settlement, slums, and upgrading • Transport, density, urban planning and urban form • Housing markets and program design • Development on the urban fringe and the city center, and the environment The remainder of this general introduction briefly describes these six thematic areas. In addition, the "section introduction" to each of these six sections in the text of the anthology will delve more deeply into these themes in order to place each paper in a useful framework. Land markets, development and policy. Legal land development for low-income households has dried up or is in the process or drying up in many developing country cities. For example, in Buenos Aires, the formal sub market for sales of individual lots in monthly installments to low-income households was important from 1950 to 1970 (World Bank, 2006), but has disappeared since then - see Box 1. During this period, land developers extended purchase-money loans to buyers (typically 150 monthly installments)-the most common form of credit finance for selling lots to low-income households in emerging countries. However, indexation of such contracts mandated by government, hyper-inflation during the 1990s eliminated these loans. Partly as s result, many subdivisions remain largely unoccupied on the fringes of Buenos Aires, and legal low- income land markets are paralyzed (World Bank, 2006). Simply adding money- either through subsidies or credit finance - without addressing such land bottlenecks results mainly in raising land prices. Put another way, the inelasticity of supply produces mainly higher prices rather than more units when demand increases. The mounting pressure on urban land has driven the rise in the price of housing, and made housing markets surrealistically dysfunctional in many major metropolitan areas of developing countries. In Dhaka, for example, the price of the median house is a startling 106 times the median annual household income. In comparison, the highest-priced metropolitan housing markets in the U.S. - New York City and San Francisco - have median-price-to-annual-household-income ratios of around 6. The extreme pressures on and high cost ofland have also lead to innovative approaches to land development that, in effect, lower the price and capture a portion of the added-value of public investment in urbanization. In particular, Asian countries - Singapore, Hong Kong, and, most recently, China - have taken measures to lower the cost basis of urban land for affordable housing and other types of development. Earlier, Japan and South Korea encouraged owners of land on the urban fringe to pool their property as a means of more efficient development - a method called "land readjustment." Some governments own considerable amounts of land in peri-urban and urban areas that is significantly under-utilized. Publicly-owned land frequently has fundamental importance for both the public and private Land and Urban Policies for Porvety Reduction 37 sectors. Typically, however, public landownership remains fragmented among many different agencies at various levels of government, each with its own mandate and administrative turf to be guarded. The ownership and legal rights to particular parcels are often in confusion. Hence, the first step usually consists of inventorying publicly-owned land along with selected privately-owned plots to clarify the legal status of these vacant or under-utilized parcels. Such investigations usually show that some parcels can be developed in a straightforward way. Other parcels are likely to have complex ownership problems that are difficult to solve in the short term. Clarifying the legal status of these parcels represents a pre-requisite for action to stimulate their use such as incentive mechanisms to place privately-owned property on the market. The first two papers of this anthology's first section explore innovative efforts to transform urban land development in order to reduce greatly its cost for affordable housing and other uses. In Iran (Keivani, Mattingly, and Majed), government limited the size of individual land holdings, resulting in transfer oflarge amounts at low cost to the public sector, which passed on these benefits to individuals and developers, and resulted in housing roughly 7% of the country's lowest-income households. Maldonado analyzes the experience of Colombia with a new legal framework for land readjustment that captures a portion of the value added by public investment in urbanization to order to finance and develop affordable serviced lots. The third paper (Pearce-Oroz) investigates the institutional realities and limits of urban land markets -which are often captured by a small elite - in the context of massive reconstruction aid after Hurricane Mitch hit Honduras in 1998. The fourth and final paper of this section documents the failure of the local property tax in Peru to produce substantial revenue due to weak local governance - a common problem in emerging countries - and the introduction of tax collecting agencies independent from municipalities that has led to great increases in property-tax revenue, although from a miniscule base. Secure tenure, property rights, and informal land delivery systems. The drying up oflegal low-income land markets leaves illegal development (variously termed "pirate'', "informal'', and "clandestine") and informal markets as the main source of land for low-income settlement, and the progressive housing process as the principal means of occupation and building of habitat in many developing country cities for low/moderate- income families. Typically, households invade land or purchase a lot in an informal sub-division and build their housing over 10 to 15 years. They finance this construction largely through their own savings, but also though many other sources including small loans, pension funds (if available), and mutual-aid arrangements with other families. As the families consolidate the house, the community lobbies for services and greater tenure security. The legal upgrading of community and the extension of services parallels the physical upgrading and building of the individual houses. Thus, progressive housing is partly an individual process - that of the house - but with a strong collective component - upgrading of services and legal status of the community. Informal land delivery mechanisms constitute parallel systems for land development and tenure .. Although these "para-legal" systems are lower cost, they are often less transparent. Again, Argentina - a middle- income, relatively sophisticated country- provides one example. Households may obtain ownership through peaceful occupation of land for 20 years, in general, and for 10 years in limited cases, and a 1994 law provides for registering the purchase agreements for such lots to increase security of tenure. This informal land system co-exists with the formal registration of property deeds. However, the cost of formal-sector registration typically ranges from US $400 to $700 including title expenses, and most low-income purchasers oflots on installments from land developers in the 1950s and 1960s have yet to sign their deeds due to lack of funds (World Bank, 2006). 38 Land and Urban Policies for Porvety Reduction These parallel informal systems also often out-compete the formal ones. In effect, the entry costs are much lower (although the total costs over time usually far exceed those of formal-sector development) and the characteristics appear better suited to the needs and effective demand of low/moderate-income households. That is, informal development typically demonstrates some combination of: (a) more central location (closer to jobs and social networks crucial to the poor); (b) larger lot size that allows poor households more room to expand and customize their habitat to their needs (larger families, home-based micro-businesses, urban agriculture); and (c) more flexible financing terms (payments can be missed if justified by temporary sickness, job loss, or other compelling causes) better suited to these household's intermittent informal incomes and employment, although interest rates are usually very high. In addition, informal land development also often benefits from the implied promise of subsequent service provision and upgrading by government, largely at public cost. These benefits get capitalized to some extent into a higher price that households pay illegal developers for a lot of raw land. "Secure tenure" of land protects these households against eviction and bull-dozing of their communities. Hence, it provides the foundation for households to invest progressively in their homes and build their communities. Full legal title backed by modern land systems (property registry, cadastre, effective legal enforcement) gives the greatest security of tenure, but is costly, technically demanding and often pushes the entry price of a~cess to the lot beyond the reach of low/ moderate-income households. In many regions, intermediate and traditional forms of property ownership have provided a sufficiently secure basis for the progressive land and housing process. Other aspects of property rights systems offer ways to address urban-land issues including: group rights vs. individual rights; and leasing/rental as opposed to ownership. Individual rights facilitate markets and transparency, but are problematic in reaching low-income households. Experiments with group rights in low-income communities - such as in Recife and Porto Alegre, Brazil and the Community Land Trusts of Kenya (see Payne) - have proved interesting, but hard to ramp up. Rental housing and long-term land leases have theoretical virtues. Long-term land leases, in principle, can offer security of tenure sufficient for financing (Deininger, 2003). Informal rental housing in poor neighborhoods already provides the main source of rental accommodation in most developing countries (Gilbert). Typically, households build an extra room or unit onto their existing home (horizontally or vertically) and rent it as a source of income. fu they do not have to pay extra for land and gain other economies (e.g. existing clandestine and legal service connections) from their adjacent owner-occupied unit, such accessory units are the least expensive way to produce low-income housing. Subsidized rental housing is the main form of affordable housing in most affluent countries. Thorny technical and political problems, however, make the expansion of formal low-income rental housing and leasing ofland difficult and rare in emerging countries. From a technical perspective, no one has solved the problem of who will own, operate, and maintain low-income rental units in a way that ensures satisfactory affordable shelter, and that channels the benefit of any p'.ublic subsidy or publicly-financed improvement largely to the low-income renters rather than mainly to the owners. Western Europe, the U.S., and Canada use networks of sophisticated non-profits and/or municipal corporations supported by public subsidy systems backed by a well-functioning legal framework to operate, maintain, and - increasingly - develop affordable rental housing. Land and Urban Policies for Porvety Reduction 39 ' ' However, most low and middle-income countries still lack such organizations and the funding and legal/ regulatory structure necessary to make this approach work, although a few are beginning to develop affordable rental systems (e.g. Singapore, Hong Kong, China). From a political perspective, most developing-country governments find production of homeowner units much more rewarding than support of rentals. In many regions - particularly in Latin America and South Asia - it could be argued that a strong cultural preference for homeownership eclipses any government effort at ren.tals, except for rent control, which generally shuts these markets down and ends up greatly reducing the stock of rental units. In contrast, the bulk of urban dwellers in some parts of sub-Saharan Africa view their urban residence as a transient place for commuting to work in the city before returning to their real homes in their tribal areas, and rental accommodations are much more common. The first paper of this section (Durand-Laserve) .documents how increasing pressures on urban land and the "commodification" of shelter and settlement has increased "market evictions" of families holding intermediate title to property, although international declarations and pressures have contributed to reducing "forced evictions." The second paper (Mooya and Cloete) uses the tools of the New Institutional Economics to analyze the argument in Hernando .De Soto's path-breaking book, The Mystery of Capital, that full legal tide is the key to turning "dead capital" in the form of informal property held by many low-income families into an economic asset and to detonating broad-based economic growth. The paper concludes that intermediate forms of tenure can have the virtues of full legal tide if properly constructed, and then examines the case of Namibia in this context. The third paper (Fernandes) documents and assesses the recent efforts of the Brazilian federal Ministry of Cities to develop a comprehensive approach for regularizing tide throughout that country. In the fourth paper, Abramo gives a structural and theoretical over-view of informal settlement in Brazil. The . fifth paper (Rakodi) looks at traditional land delivery systems in five medium-sized Sub-Saharan African cities, and concludes that policies and programs can build on their strengths. Informal Settlement, Slums, and Upgrading Although progressive housing is a crucial solution, it is also an immense problem that exacts enormous public and private costs when unguided. Increasingly, tight land markets force households to settle on precarious locations including ravines, steep hillsides, marshes, riverbanks, garbage dumps, watersheds, sidewalks, the edges of public facilities and infrastructure lines and associated rights-of-way, and distant sites far from existing infrastructure lines that are often environmentally fragile or inappropriate. Alternatively, these families crowd into ever-denser existing informal settlements: inner-city tenement units divided into many rooms with each rented to a family; and shantytowns on the urban fringe and beyond that expand horizontally into every free space and then vertically by adding stories to existing structures Slum upgrading involves retrofitting these areas with infrastructure to create a viable road network underlain by water lines, and accompanied by drainage and sanitation. This process often requires relocating a modest share ofa slum's population (around 5%) -frequently, a problematic and costly step. Slum upgrading frequently occurs piecemeal and without an overall plan or layout, mainly close to election time when candidates for political office trade an improvement or commitment for an improvement for votes. In contrast, "integrated" slum upgrading programs provide the missing basic services together based on a plan, and- often - join them with organized community participation and selected social and economic services and with legal tenure. For 40 Land and Urban Policies for Porvety Reduction these reasons, retrofitting these areas through slum upgrading is usually much more expensive than new formal-sector development. Government typically ends up absorbing the high capital costs of improving or replacing the infrastructure of these communities, selective resettlement, and regularizing their legal situation. The relatively high costs of slum upgrading have created problems for financial sustainability and program scale. Particularly when an integrated approach is taken that lifts these areas to standards approaching (buy still below) those of the rest of the city, the high cost per household tends to make these programs into boutique, small-scale efforts. The model project looks good, but cannot be expanded much. In addition to the public costs of upgrading programs, informal housing development also has high costs for families. The process of home construction is typically long and wasteful. One market study (see Box 3) found that building a basic 2-bedroom house takes Mexican families 11 years, and costs 30% more because of the high cost of small purchases of building materials, theft and damage of these materials, and poor planning. Households also end up paying high sums for purchasing a raw lot, for improving security of tenure, for basic services (e.g. private water supplied by tanker, which is typically 5 to 10 times the cost of publicly-supplied water) prior to consolidation, and to save and to borrow sums for the steps in the progressive housing process. Irregularly-settled neighborhoods also have substantially higher levels of crime and insecurity than other neighborhoods of a similar socio-economic profile. The bad reputation of these neighborhoods can brand their ~esidents, and make them largely unemployable in the formal sector (e.g. Jamaica). The high public and private costs of upgrading existing slums have called attention to the importance of slowing the formation of new slums through getting ahead of demand by expanding low-income land development. This strategy holds particular importance in South Asia and Africa where urbanization is still cresting. Most medium and large developing-country cities are still growing at rates that will double their size in 20 to 25 years. The global population is projected to increase by 1.5 to 2 billion over this period, and the bulk of these people will constitute low-income households living in developing country cities. Where will all these new city residents live? As Payne notes, the international community has come to realize that the "real challenge of slums is two-fold:" First, there is a need to improve the living conditions of people living in slums and various types of unauthorized settlements. And second, there is an equally urgent need to create conditions in which all sections of urban society, especially the poorest and most vulnerable, can obtain access to legal, affordable shelter in ways that prevent the need for future slums and unauthorized settlement. In the first paper of this section, Abiko, Azevedo, Reinaldelli, and Haga quantify slum upgrading costs in Brazil, and find that providing a basic package of services costs through these programs costs around three times (US $3,000) that of formal-sector development (US $1,000) on average, although these costs range widely between simple and complex projects. The second and third papers show that the likelihood that certain areas will become slums and that households will become slum dwellers can be predicted, and- thus - that proactive advance planning can have a large impact on meeting the challenges of slums. Sietchiping's application of a Geographic Information System based on mathematical "cellular automata" dynamically maps urban development in Yaounde (Cameroon in West Africa) and predicts the location of slums with 73% accuracy. Piedade, Oliveira, and Albuquerque use a probit model to determine the likelihood that Brazilian households with specific socio-economic characteristics (higher unemployment, lower quality of employment, lower schooling, higher household sizes, etc.) will live in a slum. The fourth and final paper Land and Urban Policies for Porvety Reduction 41 (Betancur) examines an integrated slum upgrading program in the context of urban violence and local politics in Medellin, Colombia. Transport, density, urban planning and urban form The immensity and paradoxes of the urban land challenge suggest that the most effective solutions must join the micro level of projects with that of the macro-development of the city region as a whole. Here, innovations in transport and urban planning, systems of settlements and the form oflarge metropolitan areas are crucial. Urban-density studies (such as density-gradient analysis) demonstrate that housing and transport are a binomial equation. Improvement in urban transport opens up much larger land areas for residential development and improves economic productivity. In turn, higher residential densities make public transport systems economically feasible. The form of metropolitan areas is crucially important to both housing and transport. This is particularly true for the immense urban agglomerations - or "megapolitan areas" - that contain an increasing share of populations - such as those of Mexico City, Sao Paulo, andJabotabek (i.e. Jakarta and surrounding areas). Based on the experience of Asian megapolitan areas, Laquian concludes, "allowing a monocentric settlement to grow in an uncontrollable and haphazard fashion is a recipe for disaster. .. (These areas are) sprawling, and extremely expensive to provide basic services." Instead, macro land-use decisions and other measures can create poly-nucleated urban region. Traditional master planning (zoning, subdivision regulation), typically leaves blank spaces for the huge informal settlements within developing country cities, and is of little use. Instead, strategic plans should focus on systems of settlements. Relatively simple actions such as laying out main roads in a rational way (Angel) in expansion areas can also have an important impact. Improving the governance and management of metropolitan regions (Freire and Stren) has crucial importance for implementing such macro approaches. However, many metropolitan regions in developing countries as developed countries are fragmented into dozens oflocal jurisdictions and authorities, and the institutions for coordination among them are only gradually emerging. The first paper (Serra, Dowall, Motta, and Donovan) examines the form of three Brazilian cities - ReCife, Curitiba, and Brasilia - through calculating population density gradients and regression analysis of the determinants of land prices. The relatively well-functioning land markets of Recife and Curitiba contrast with those of Brasilia, and raise important issues for social welfare and economic development. Coelho and Irving continue this type of analysis by calculating density gradients for 10 Brazilian cities. The third paper (Graham) concerns the links between city size, productivity, and infrastructure provision through calculating elasticities of productivity with respect to city size for different industrial sectors in the United Kingdom. . Kumarage then examines the impact of transport investment on urban poverty and land development in Colombo, Sri Lanka, concluding that improving transport holds key importance for low-income households. The final paper (Pujol) analyzes the metropolitan development of San Jose, Costa Rica. While Costa Rica has, in significant measure, met its housing challenge, many urban-development issues remain. Housing Markets and Low-Income Housing Programs Starting in the early 1990s, many governments and donors - influenced by the World Bank - adopted an "enabling markets" approach.to housing (World Bank, 1993). The context of the emergence of this approach 42 Land and Urban Policies for Porvety Reduction consisted of the fall of the Soviet Union and entry into the market system of a large share of the world's population (in China, India, and the Newly-Independent States), the poor results of highly-subsidized housing programs that attempted to replace the market in many countries - particularly in Latin America, the limited impact of sites and services and slum upgrading projects, and the Savings and Loan debacle of the 1980s in the U.S., where ignoring the logic of markets cost taxpayers US $500 billion. The enabling-markets approach has encouraged reform of various aspects (land, property rights, infrastructure, housing finance, housing institutions) of the housing "sector", and embraced land issues within a housing framework. This approach lead the World Bank to shift from supporting sites and services and slum upgrading - which were viewed as isolated projects with little systemic impact - to reforming and expanding mortgage credit in the hope of eventually pushing this and other aspects of formal-sector housing "downmarket" to reach low/moderate-income households. Enabling housing markets has had a number of successes. "In particular, mortgage finance -which was formerly available mainly in OECD countries - has now spread throughout the world (Buckley, 2005). However, formal systems - including mortgage credit - have largely failed to go downmarket to reach low-income households. In most countries, even moderate-income families remain left out of formal-sector housing and land systems. Meanwhile, the "informal sector" and slums - which appeared a limited market failure in the early 1990s - have continued to grow in many regions. In Sub-Saharan Africa, where many countries have urbanized rapidly without economic growth (Fay), these irregular settlements consume the great bulk of many cities. It is now clear that these impoverished, poorly housed, and poorly serviced areas are at least semi- permanent features of the urban landscape in many regions. In retrospect, the initial enabling-markets approach appears too sanguine about the difficulties of creating "well-functioning" housing markets - where "everyone is housed adequately ..... at a reasonable share of income" and "residential land is available at a reasonable price" (World Bank, 1993). The urban process is also much more complex and diverse now than when the World Bank first started its work (Buckley). Well- functioning housing and land markets are powerful but difficult to create and maintain, and must frequently be supplemented with interventions to overcome large-scale market failures. This is not only true in developing countries but also, arguably, in affluent countries. Housing affordability has sharply declined in Western Europe and the U.S. in recent years. 1 Some (Laquian) have speculated that the "enabling markets" approach appears to be a "transition to a moment when much greater and more systematic attention needs to be paid to housing, land, and urban development." This is not just the job of the public sector. GDPs of developing countries as a whole are growing at over 6% per annum, compared to rates of around 2% for the developed world. Housing is the largest single investment of the low/moderate-income majority. Surely, if markets are to play a substantiaT role in development, then the private sector could have a substantial role in low-income housing and land. However, the private-sector organizations that employ the most effective 1 Although both the U.S. and much of Western Europe are in the middle of a housing market correction, housing affordability has declined steadily for half a century, particularly in large metropolitan areas. For example, the median house price-to-median household income for the U.S. as a whole has gone from 2 in the 1950s to over 3 today. Only more favorable finance (some aspects of which create greatly increased risks for households and for the financial system) and two-income families have kept the rate of homeownership from falling in the U.S. Land and Urban Policies for Porvety Reduction 43 management methods and that have the greatest capacity to help low-income households - multi-nationals and large local companies - generally do not understand low-income markets, and - with some notable exceptions (Prahalad) - have kept out of them. Instead, marginal producers and suppliers of land, building materials, finance and other inputs to the land/housing process dominate. The result is, too often, very high- cost, "savage" low-income housing and land markets (Buckley) in which local bosses and public and private mafias greatly increase costs at transition points. Thus, the methods and models for involving the private-sector constructively in solving low-income housing problems largely remain pending. An encouraging exception is that of CEMEX, the third largest cement maker in the world, in satisfying markets for progressive housing in Mexico .. The CEMEX Patrimonio Hoy program organizes small groups of families who commit to a 70-week saving program, arranges with local building materials suppliers to deliver high-quality product at competitive prices, and advances microcredit to these families in the form of delivering building materials well prior to payment by households. CEMEX operates this program through establishing offices located in low-income communities, and local "promoters" - 98 percent of them women - to inform local households about the program. Patrimonio Hoy has proved astonishingly successful. reaching 100,000 people in its first two years with plans to expand this number to 1,000,000 in the next 5 years. The program operates without subsidies and the other two of the top three cement manufacturers of the world - Folcin and Lafarge - have recently launched initiatives to reach the progressive housing market in a number of developing countries. Hence, the involvement oflarge corporations and application of modern management methods to low-incoming housing still has potential, despite the uneven results of a decade and a half of "enabling housing markets." Due to the crucial importance of urban land for the poor and the failure of the enabling markets approach to address this problem, a land-centered approach appears to be replacing a housing-centered approach to low- income shelter and settlement. Nevertheless, the traditional issues of housing finance - including how to join housing credit, housing savings systems, and housing subsidies to make shelter more affordable - remain. On the real side, an important area for innovation and program design is the various forms of "low-cost housing solutions." "Low-cost land and housing solutions" consist of a wide range of options that compose the steps of the progressive housing process. These include serviced and unserviced lots, rehabilitation and improvement, expansi~n, construction of a core unit on a lot already owned by the family (for replacement, to add a unit, for rental), tenure regularization, infrastructure and service upgrading etc). These incremental housing solutions cost a small fraction of purchasing of a new commercially-built unit. Thus, they represent a fundamental key to large-scale provision of affordable shelter and housing policy in many countries. Joining such project approaches with new technologies including housing microfinance (Ferguson), organized community participation (Ruster and Imparato), and selective involvement of the private sector- such as the Patrimonio Hoy program of CEMEX - may hold the key to creating a new generation of more effective, piore sustainable, and more massive low-income housing projects that really,,_do reach the poor at scale. In this context, it may be time (Buckley) to re-evaluate the earlier experience of the World Bank and county governments with sites and services, and slum upgrading programs. In contrast, many government housing programs still often focus on making moderate and middle-income families bankable in order to move formal-sector credit and other systems downmarket to these groups and to 44 Land and Urban Policies for Porvety Reduction spur economic growth. Physically, the prototype moderate-income housing solution in Latin America consists of a core expandable unit of 25 to 45 m2 that families upgrade and expand in programmed steps, as need and· available resources dictate; and, in East Asia, a 40 to 80 m2 unit in a multi-storey building. The vested interests of the construction and development industry often play a large role in promoting this policy approach. However, most developing countries usually have a very small housing credit system and a potentially more important instrument is subsidies (Buckley). The art of low-income housing program design consists mainly of joining financial resources (subsidies, credit, and household savings) with different types of low-in,come housing solutions to suit local housing conditions, the financial capacity of government to fund these efforts, and the institutional capacity of other key actors (housing NGOs, local governments, lenders) to perform their roles in these efforts. Within the developing world, the housing programs of Latin America, and those of East Asia are particularly noteworthy. While Latin America has focused on housing subsidy systems, East Asia has emphasized forceful public management of urban land. The fi~st two papers of this section examine housing programs in Latin America. Zanetta examines the decentralization of Argentina's National Housing Fund to provincial governments during the 1990s, while Fonseca, Trani, and Wakisaka document the large effort of the state of Sao Paulo in affordable housing, in general, and the experience of its self-help housing partnership with the state's municipalities, in particular. The third paper adapts the model used in the state of Florida in the U.S. for estimating housing need to Brazil. The last two papers deal with East Asia. Yuen examines the strikingly successful experience of Singapore, and Zhu surveys that of Singapore, China, Bangladesh, and Vietnam. The most successful Asian cases - Singapore, Hong Kong, and China - join forceful public management or ownership of urban land with private-sector development and ownership of the resulting units. Development on the urban fringe, the city center, and the environment Development on the urban fringe increasingly takes polarized forms in developing countries. Low-income households - although not the poorest, whose main priority is to locate close to jobs in the city center - tend to occupy sprawling informal subdivisions on the periphery. Subsidized government housing development for low/moderate-income families depends on the availability oflow-cost land, also located on or beyond the urban fringe. At the other end of the income spectrum, the elite follow manufacturing subsidiaries of international companies (Buenos Aires, Curitiba, Sao Paulo), universities, locally-grown high-tech manufacturers and international service-providers (Bangalore and some other Indian cities), and commercial establishments to the suburbs, and increasingly live in gated suburban communities. The resulting sprawl has strong negative environmental impacts. It consumes agricultural and environmentally- sensitive land. City growth also contributes to threatening an absolute global shortage of fresh water. Utility companies must go further and further afield to find sources, and spend skyrocketing sums on processing, pumping, and transporting it. De-salinization technologies may have a role to play here in coastal cities. Most troubling of all, sprawl joined with the export of old, highly-polluting automobile technologies from rich countries to China has substantially increased the carbon-dioxide emissions of the world, and contributed to global warming. The alternative to sprawl involves grc;ater densification of existing urban areas, particularly around transport nodes. In this regard, many larger and older developing-country metropolitan areas have come to assume the Land and Urban Policies for Porvety Reduction 45 donut form of U.S. cities. Congestion, crime, and departure of the middle-class to the suburbs leave a declining central city, which sometimes falls in population. Redevelopment of central cities appears to make sense. After all, these areas already have infrastructure and services, and are much closer to jobs than the periphery. On closer inspection, however, the costs of purchasing, cleaning (necessary for "brownfields" sites formerly used for polluting industries), and developing centrally located sites are usually higher than development on the fringe. Strong public-private partnerships are essential to assemble sufficiently-large parcels of centrally located land to make such redevelopment projects economically viable. While redevelopment of central cities has a long history in the U.S. and Western Europe, most developing countries are only now beginning to build the institutions and legal framework for such partnerships. The first four papers of this final section of the anthology examine the dilemmas of rapid development on the urban fringe - the pattern of most developing-country cities. The last two papers look at the theoretical advantages and practical difficulties of redevelopment of the central city and the densification of existing urban areas. Pantelic, Srdanovic, and Greene note that the distinct features of the urbanization of the past two decades constitute a "post-modern period" in which the segregation of the rich and poor increasingly makes low-income urban households vulnerable to natural and man-made disasters. Sridhar then examines suburbanization of newly-prosperous Indian cities. The third paper (Torres) analyzes the urban sprawl of the Sao Paulo metropolitan area, the expansion of its poor periphery, and the impact on the environment. The fourth paper of this section (Goytia) focuses on the polarization of development of the urban fringe into gated communities for the rich and slums for the poor in Pilar, a municipality in the northeast corner of the Buenos Aires metropolitan area. Aragao then investigates the central-city redevelopment experience of Paris and Barcelona, and compares it with that of Sao Paulo. The final paper of this anthology examines the densification of Guadalajara, Mexico's second largest city, concluding that the methods to promote a more compact city are still incipient in this metropolitan area. References Angel, Shlomo. 2006. Preparing for Urban Expansion in Intermediate Cities in Ecuador. Draft Paper prepared for the World Bank. Task manager: Alexandra Ortiz. Buckley, Robert. 2005. Thirty Years of World Bank Shelter Lending. Washington, D.C.: World Bank. Deininger, Klaus. 2003. Land Policies for Growth and Poverty Reduction. Washington, D.C.: World Bank. Durand-Laserve, Alain and Lauren Royston (eds). 2002. Holding Their Ground. London: Earthscan Publications. Fay, Marianne and Charlotte Opal. 2000. Urbanization without Growth; a not-so-uncommon phenomenon. Research Working Paper no. WPS 2412. Washington, D.C.: World Bank Ferguson, Bruce. "Chapter 2- The Key Importance of Housing Microfinance" in Daphnis, Franck and Bruce Ferguson (eds). in 2004. Housing Microfinance; A Guide to Practice. Bloomfield, CN: Kumerian Press. Freire, Mila and Richard Stren (eds). 2001. The Challenge of Urban Governance. Washington, D.C.: World Bank. 46 Land and Urban Policies for Porvety Reduction Gilbert, Alan. 2003. Rental housing; an essential option for the urban poor in developing countries. New York: United Nations Human Settlements Program. Laquian, Aprodicio. 2005. Beyond Metropolis. Washington, D.C.: Woodrow Wilson Center Press. Payne, Geoffrey (ed). 2002. Land, Rights & Innovation. London: ITDG Publishing. Prahalad, C.K. 2005. The Fortune at the Bottom of the Pyramid. Upper Saddle River, N.].: Wharton School Publishing. Ruster, Jeff and Iva Imparato. 2003. Slum upgrading and participation: lessons from Latin America, Volume 1. Washington, D.C.: World Bank. World Bank. 2006. Review of Argentina's Housing Sector. Team leader: Maryse Gautier. Draft document. Washington, D.C. World Bank. 1993. Enabling Housing Markets. Principal authors: Steve Mayo and Shlomo Angel. Washing- ton, D.C.: World Bank. INTRODU~AO GERAL SIMPOSIO INTERNACIONAL SOBRE PESQUISA URBANA 2005 Mila Freire, Bruce W Ferguson, Ricardo Lima, Dean Cira and Christine Kessides Como observou Sir Peter Hall em seu prefacio, a "Revolw;:ao Urbana'' verificada em especial nos pafses em desenvolvimento representa grandes oportunidades e riscos. A urbanizac;:ao tern a faculdade de ajudar na eleva- c;:ao dos padroes de vida, pode oferecer a infra-estrutura e os servic;:os que resultam em enorme aprimoramento em termos de bem estar humano, alem de libertar as pessoas do jugo da terra e do domfnio total da busca quotidiana por alimentos. Os bairros e centros atraentes, 0 transporte eficiente, as varias amenidades, OS impres- sionantes indicadores sociais e o elevado padrao de vida de Cingapura e de Curitiba sfo sinais desse potencial. Por outro lado, caso seja mal administrada, a onda urbana podera resultar em acentuado crescimento da pobreza nas cidades, gerando condic;:oes surrealisticamente desesperadoras e fomentando a doenc;:a e a violen- cia. Os moradores de rua de Bombaim, que convivem lado a lado com a imensa riqueza desse capit6lio comercial de uma India que vem recentemente conhecendo a prosperidade, e os casebres e favelas aparente- mente sem fim que consumem tantas cidades da Africa subsaariana sfo exemplos emblematicos desse outro presente urbano e de seu poss!vel futuro. No centro de muitas dessas oportunidades e riscos, esra o pr6prio terreno urbano. Encontrar lores de terra bem localizados e a prec;:os razoaveis e hoje 0 desafio mais crucial para 0 desenvolvimento de habitac;:oes acess!veis. Nos casos em que esse tipo de programas nao estejam dispon!veis - coma ocorre com freqiiencia- a maioria formada por pessoas de renda baixa a moderada em muitas cidades de pafses em desenvolvimento em geral nfo consegue pagar por uma casa mais barata constru!da comercialmente, recorrendo em vez disso aos sistemas "informais" para solucionar o pr6prio problema habitacional. Mesmo esse tipo de "habitac;:ao progressiva'' inicia-se com o acesso a um lote, dependendo tambem dele. Da mesma forma, a efici~ncia do transporte e a capacidade das famllias de garantirem a conexao com seus empregos e servic;:os estao condicio- nados ao USO e a densidade da terra. No pedodo que caracterizou o in!cio da grande onda urbana nos pafses em desenvolvimento - a decada de 1950 - as fam!lias pobres que migravam dos campos para as cidades conseguiam com alguma freqiiencia encontrar terrenos baratos e de localizac;:ao central onde instalar-se. 0 filme "Orfeu Negro", que recria o mito de Orfeu e Euddice nas favelas localizadas em morros fngremes e com vistas panodmic;:as da cidade do Rio de Janeiro, pinta um quadro idllico da vida na favela na epoca. Hoje em dia, e impossfvel imaginar um filme lfrico coma esse sendo produzido sobre a vida que ali se desenrola. Na verdade, atualmente o cinema brasilei- ro produz muitos filmes neo-realistas asperos, que descrevem os resultados das condic;:oes cada vez piores e mais espalhadas das favelas, incluindo os 6rfaos de rua, seqiiestros e violencia urbana. Em relac;:ao a isso, na maior parte das cidades de pafses em desenvolvimento a era de acesso facil a terrenos urbanos ha muito tempo acabou, uma vez que o avanc;:o da urbanizac;:ao terminou por ocupar as melhores areas para desenvolvimento ao redor de muitas cidades. Muito embora alguns terrenos localizados em setores 48 Land and Urban Policies for Porvety Reduction urbanos e periurbanos sejam freqiientemente de propriedade de 6rgaos dos governos, as grandes empresas incorporadoras, que constroem principalmente visando fam.flias de renda alta e media, agora parecem ser donas da maior parte dos lotes adequados a constrw;:ao que ainda restam. Os valores reduzidos, os altos requisitos tecnicos e as dificuldades pol.fticas dos impastos sobre a propriedade imobiliaria nos pa!ses em desenvolvimento permitem que esses grandes proprietarios de terras continuem mantendo os terrenos, incor- rendo em custos reduzidos. Na ausencia de medidas que venham a melhorar essa situac;:ao, de processos que formalizem a propriedade da terra e de outras reformas, o resultado vem sendo a "commodificar;ao" da terra e da habitac;:ao em geral (Durand-Laserve), o que com freqiiencia leva a elevac;:ao dos prec;:os ea exdusao dos pobres. Por muitos motivos, a terra urbana tornou-se hoje a principal "restric;:ao vinculante" a questao habitacional relacionada a populac;:ao carente. A presente antologia reline e organiza 32 trabalhos apresentados durante o Simp6sio Internacional de Pesqui- sa Urbana, realizado em Bras.flia de 4 a 6 de abril de 2005, sob o patrodnio conjunto do IPEA e do Banco Mundial, e que focalizou a terra urbana. " Os trabalhos apresentados durante esse Simp6sio de 2005 foram organizados ao redor de seis areas tematicas nas sec;:oes desta antologia: • Mercados e desenvolvimento da terra e pol.ftica do solo • Garantia de posse, direitos de propriedade e sistemas informais de acesso a terra • Assentamentos informais, favelas e urbanizac;:ao • Transportes, densidade, planejamento urbano e conformac;:ao urbana • Mercados de im6veis residenciais e desenho de programas • Desenvolvimento na periferia urbana e nos centros das cidades e o meio ambiente A seguir, nesta introduc;:ao geral, passaremos a descrever brevemente essas seis areas tematicas. Ainda, a "intro- duc;:ao da sec;:ao" qu~ antecede cada uma das seis sec;:6es constantes do texto da antologia tratara em maior detalhe dos t6picos, de forma a colocar cada um dos trabalhos em seu contexto. Mercados e desenvolvimento da terra e poli'.tica do solo. Em muitas cidades de pa.fses em desenvolvimento, o desenvolvimento de terrenos destinados a habitac;:6es que venham a atender fam.flias de baixa renda legal- mente esta estagnado, ou em processo de estagnar-se. Por exemplo, em Buenos Aires o submercado formal representado pela venda de lotes individuais a fam.flias de baixa renda atraves de prestac;:oes mensais foi impor- tante de 1950 a 1970 (Banco Mundial, 2006), mas desapareceu desde entao - ver Quadro 1. Durante esse per.fodo, os incorporadores forneceram diretamente aos compradores emprestimos para a aquisic;:ao dos im6- veis, que ficavam garantidos por gravames sobre os mesmos (tipicamente, 150 prestac;:6es mensais) - a forma mais comum de credito para a venda de lotes a fam.flias de baixa renda nos pa.fses emergentes. Entretanto, a indexac;:ao desses contratos exigida pelo governo ea hiperinflac;:ao dos anos 90 terminaram por impossibilitar esses emprestimos. Em parte como resultado disso, muitos bairros na periferia de Buenos Aites permanecem em grande parte desocupados, enquanto estfo paralisados OS mercados legais de terra destinada a populac;:ao de baixa renda. (Banco Mun dial, 2006). Simplesmente acrescentar dinheiro - quer seja atraves de. subs.fdios ou credito para financiamento - sem solucionar esses gargalos relacionados a terra principalmente tern como resultadci a elevac;:ao dos prec;:os dessa Land and Urban Policies for Porvety Reduction 49 terra. Em outras palavras, a falta de elasticidade da oferta produz principalmente prec;os mais altos, em vez de um num~ro maior de unidades, quando a demanda cresce. A pressao crescente sabre o solo urbano determinou o aumento dos prec;os dos im6veis residenciais, tendo tornado o mercado surrealisticamente disfuncional em muitas das principais areas metropolitanas dos pafses em desenvolvimento. Em Dacca, por exemplo, o prec;o de uma casa media atinge o nfvel chocante de 106 vezes a renda anual familiar mediana. Em comparac;ao, os mercados de residencias metropolitanas mais caros nos Estados Unidos - os da cidade de Nova Iorque e de Sao Francisco - exibem proporcr6es entre os prec;os medias e OS rendimentos anuais medias das famflias que correspondem a aproximadamente 6. As press6es extremas sabre o solo e o seu alto custo geraram ainda abordagens inovadoras ao desenvolvimento da terra, que efetivamente reduzem o precro e capturam .uma porc;ao do valor agregado do investimento publico em urbanizac;ao. Em especial, os pafses asiaticos - Cingapura, Hong Kong e mais recentemente a China - tomaram medidas para reduzir a base de custo dos terrenos urbanos para residencias acessfveis e para outros tipos de desenvolvimento. Ainda antes disso, o Japao ea Coreia do Sul ja estimulavam os proprierarios de terras nas periferias das cidades a reunirem suas propriedades, coma forma de promover um desenvolvi- mento mais eficiente - um metodo denominado "reajuste da terra''. Alguns governos sfo OS proprietarios de extensoes consideraveis de terras nas areas periurbanas e urbanas, que sao significativamente subutilizadas, ainda que freqi.ientemente essas terras publicas tenham imporrancia fundamental tanto para o setor publico quanta para o privado. Por outro lado, o que e mais tfpico e que a propriedade do poder publico sabre a terra fique fragmentada entre varios 6rgaos de diferentes nfveis de governo, cada um com o seu pr6prio mandato e sec;ao administrativa a ser guardada. E comum haver confu- sao entre a propriedade e os direitos legais sabre determinados lores; assim, o primeiro passo em geral envolve · a realizac;ao de um inventario da terra publica, juntamente com terrenos de particulares, para que se possa esclarecer a situac;ao legal desses lores vazios ou subutilizados. Essas investigac;6es comumente revelam que alguns lores estao promos para desenvolvimento de forma bem direta, enquanto que outros envolvem proble- mas complicados de propriedade, que dificilmente podem ser resolvidos a curto prazo. A determinacrao da situac;ao legal desses terrenos e pre-requisito para qualquer tipo de ac;ao que possa estimular o seu uso coino mecanismos de incentivo para a colocac;ao no mercado de terras particulares. Os dais primeironrabalhos da primeira sec;ao desta antologia exploram os esforc;os inovadores de transforma- crao dos desenvolvimentos em terras urbanas, com a finalidade de reduzir em muito OS seus CUStOS para construc;ao de residencias acessfveis e outros usos. No Ira (Keivani, Mattingly e Majed), o governo limitou a extensao das terras de propriedade individual, o que resultou na transferencia para o setor publico de grandes areas a baixo custo, beneffcios que foram repassados a indivfduos e incorporadores, o que por sua vez resultou na oferta de habitac;6es para aproximadamente 7% das famflias de mais baixa renda no pafs. Maldonado analisa a experiencia da Colombia, com uma nova estrutura jurfdica para o reajuste de terras, que captura uma parte do valor agregado por investimentos publicos em urbanizac;ao para financiamento e desenvolvimentos de lotes com disponibilidade de servic;os ea precros acessfveis. 0 terceiro trabalho (Pearce-Oroz) investiga as realidades institucionais e os limites dos mercados de terras urbanas - que com freqi.iencia sao dominados por uma reduzida elite- no contexto da vultosa assistencia para reconstruc;ao que se seguiu apassagem do furacao Mitch par Honduras, em 1998. 0 quarto e ultimo estudo dessa secrao documenta o Fracasso dos impastos locais sabre a propriedade no Peru em produzir um volume substancial de receitas, coma resultado das fragili- dades da governancra local - problema comum em pafses emergentes - e a introducrao de 6rgaos para a 50 Land and Urban Policies for Porvety Reduction arrecada<;:ao .de impastos independentes dos munidpios, que resultou em grandes aumentos na receita gerada par impastos sabre a propriedade, apesar da base reduzidissima. Garantia de posse, direitos de propriedade e sistemas informais de acesso aterra. A estagna<;:ao dos merca- dos legais de terra para a popula<;:ao de baixa renda faz com que os desenvolvimentos ilegais (que recebem varias denomina<;:6es, coma "piratas", "informais" e ''clandestinos") e os mercados informais passem a ser a principal fonte de terrenos para assentamentos das familias carentes. Aqui, o processo de habita<;:ao progressi- va e o principal meio de ocupa<;:ao e constru<;:ao de moradia para familias de renda baixa ou moderada em muitas cidades de paises em desenvolvimento. Tipicamente, as familias invadem a terra ou adquirem lotes em um bairro informal, construindo suas casas ao longo de dez a quinze anos. Financiam a constru<;:ao em grande parte atraves da sua pr6pria poupan<;:a, mas tambem atraves de muitas outras fontes, que incluem pequenos emprestimos, fundos de pensao (se houver disponibilidade) e arranjos que envolvem assistencia mutua entre outras familias. A medida que as familias vao consolidando a sua casa, a comunidade passa a demandar servi<;:os e uma maior garantia de posse. As melhorias legais acrescentadas as comunidades ea expansao dos servi<;:os acompanha a urbaniza<;:ao fisica e a constru<;:ao de moradias individuais. Assim sendo, a habita<;:ao progressiva revela-se em parte um processo individual - o da pr6pria habita<;:ao - ainda que inclua um forte componente coletivo - a melhoria dos servi<;:os e a condi<;:ao de legalidade da comunidade. Os mecanismos informais de acesso a terra constituem sistemas paralelos para desenvOlvimento e posse da terra. Muito embora esses sistemas "paralegais" envolvam custos mais baixos, com freqiiencia sao menos transparentes. Mais uma vez, um exemplo ea Argentina - pais de renda media e relativamente sofisticado. As familias conseguem obter a propriedade atraves da ocupa<;:ao padfica da terra durante vinte anos, de forma geral, e durante dez anos em um numero limitado de circunsrancias; uma lei de 1994 regula o registro dos acordos de compra i1a------------------------I -.-Brasilia ...... 6 100 l--~-----""'-----------------------1 ·.;::; ~Curitiba 3 a. 80 l---~:-------"i!!>f;l~-----------------1 -&-Recife ~ 60 1------=-'ilt"---------'"""t;bo;,--------------1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Distance from City Center (km). Figure 1 illustrates the population density functions for the three cities for the year 2000. Recife's gradient has the highest intercept, at 165 persons per hectare, and has a relatively flat but negative gradient of -7.3%, indicating that its population density declines at 7.3% per kilometer. Curitiba, has a lower intercept value - 124 persons per hectare - and a steeper negative gradient of -16.6%. This is interesting since Curitiba is regarded as having a very efficient transportation system - suggesting that its gradient should be flatter. However, incomes and wages in Curitiba are higher than in Recife and therefore travel time costs are higher. Land and Urban Policies for Porvety Reduction 75 Brasilia's very low population density intercept of 15 persons per hectare reflects the fact that the center of Brasilia is comprised of nonresidential activities and open spaces. Its density gradient is positive, indicating that population density increases with distance from the center. Formal residential areas are located outside the central city area, and informal residential developments are located well beyond the capital's center. Housing Stock Trends in the Three Cities During the 1990s, all three cities produced substantial housing units. In order to tentatively gauge the size of the housing stock- both informal and formal - the study relies on two sources of data: (1) the IBGE 1991 and 2000 Censuses and (2) detailed local studies on estimations of the housing sector in the three cities. According to local estimates, Brasilia added 167,682 units between 1991 and 2000 - a 46.8% increase. In Curitiba, the housing stock increased by 199,655 units between 1991 and 2000 - a 37.4% increase. In Recife, the housing stock grew by 35.1%between1991and2000, registering a 234,240 unit increase. With the exception of Brasilia, housing stock growth rates substantially exceeded the rate of increase in the population It is common for the housing stock to grow at a faster pace than population when household size is falling and when the housing market is responding to a bacldog of unmet demand. In the long run housing stock growth should closely match the rate of increase in household formati~n. Table 4: Formal and Informal Housing Stock in Brasilia, Curitiba, and Recife Metropolitan Regions, 1991-2000 BRASILIA 1991 - 2000 Absolute Increase 167,682 units, Annual Avg Increase 18,631 units, CAGR 4.4% TYPE OF HOUSING 1991 % SHARE 2000 % SHARE Formal 351,803 98.4 482, 189 91.8 Informal 5,836 1.6 43, 132 8.2 Total 357,639 100.0 525,321 100.0 CURITIBA 1991 - 2000 Absolute Increase 199,655 units, Annual Avg l.ncrease 22, 184 units, CAGR 3.6% TYPE OF HOUSING 1991 % SHARE 2000 % SHARE Formal 499,062 93.6 684,891 93.5 Informal 34, 110 6.4 47,936 6.5 Total 533, 172 100.0 732,827 100.0 RECIFE TYPE OF HOUSING 1991* % SHARE 2000 % SHARE Formal 667,818 74.0 Informal 234,721 26.0 Total 902,539 100.0 * The absence of data for 1991 is explained by the city's absence of a cadastre for the metropolitan area. In the mid 1990s, a modernized cadastre was developed, which was used to calculate the 2000 data. Flavio de Souza, of the Department of Architecture and Urbanism of the Universidade Federal de Alagoas, cited research.from a 1993 study by the Secretaria do Planejamento e Meio Ambiente calculating that 30.6% of the housing stock in Recife was informal. · 76 Land and Urban Policies for Porvety Reduction In Brazil, as in other countries, it is useful to differentiate between formally and informally provided housing. Formal provision refers to housing development that is located on legally subdivided and permitted land, where there is clear title to properties. The design of the subdivision and the housing units follows all government regulations and standards. Informal housing, on the other hand, refers to housing development that does not follow government regulations and standards and is frequently on lands that are illegally subdivided or occupied. Table 4 provides a breakdown of housing production for the three cities into formal and informal categories. Formal housing makes up the majority of the housing stock in the three cities, ranging from 92% of the stock in Brasllia, to 94% in Curitiba, and 74% in Recife as of 2000. However, in the case of Brasilia and Curitiba, there is evidence that the portion of informally produced housing is increasing. Informal housing production in Brasilia increased by 639.l % between 1991 and 2000, a compound annual increase of 22.9% (versus 3.6% per year for formal housing). This dramatic difference in growth rates drove up the portion of total informal housing stock- from 1.6% in 1991 to 8.2% in 2000. In Curitiba, the pattern is similar. Between 1991 and 2000, the informal housing stock increased by 205%, a compound annual increase of 8.3%. Formal housing stock increased by 87.5%, a compound annual increase of 7.2%. As a result, Curitiba's share of informal housing increased from 7.4% to 1.5% over the 1991-2000 period. A Closer Look: Spatial Patterns of Population, Urban Development, Population Density and Housing in the Three Cities In this section, we examine the spatial structure of the three cities, looking at the distribution of population, the compactness of urban land development in terms of population and housing, and urban land use, which provides the opportunity to compare and contrast the overall compactness of urban development. We measure compactness by calculating the cumulative percentage of total population located within specific radii of the city center. Compactness will change over time depending on the spatial distribution of residential development taJ<:ing place between 1991 and 2000. Tables 5, 6 and 7 array the spatial distribution of population for the three cities for 1991, 2000 and change between 1991-2000 according to seven distance bands, expressed in terms of distance (kilometers) from the city center. In order to foster comparison, the bands are defined to reflect the overall spatial distribution of the three cities. Table 5: Spatial Distribution of Population: Brasflia, Curitiba and Recife, 1991 DISTANCE BRASILIA CURITIBA RECIFE CATEGORY (KM) POPULATION % OF TOTAL POPULATION % OF TOTAL POPULATION % OF TOTAL 0-5 4,525 0.3 466,467 22.7 335,685 11.5 5.1-10 118,395 8.9 963,747 47.0 106,250 36.7 10.1-15 114, 125 8.6 269,572 13.2 740,296 25.4 15.1-20 214,030 16.1 174, 146 8.5 372,413 12.8 20.1-25 275,331 20.7 57,633 2.8 151,707 5.2 25.1-30 357,021 26.8 106,449 5.2 601,47 2.1 30+ 248,991 18.7 12,780 0.6 187,685 6.4 Total 1,332,418 100.0 2,050,792 100.0 2,917,183 100.0 Land and Urban Policies for Porvety Reduction 77 Table 6: Spatial Distribution of Population: Brasf\ia, Curitiba and Recife, 2000 DISTANCE BRASILIA CURITIBA RECIFE CATEGORY (KM) POPULATION % OF TOTAL POPULATION % OF TOTAL POPULATION % OF TOTAL 0-5 6039 0.3 480,872 18.5 344,205 10.3 5.1-10 152,212 7.4 1,051,713 40.5 1146,924 34.3 10.1-15 140,754 6.8 422,786 16.3 869, 114 26.0 15.1-20 334,091 16.5 296, 169 11.4 488,738 14.6 20.1-25 319,336 15.5 120, 767 4.7 183,384 5.5 25.1-30 497,216 24.2 193,643 7.5 78,566 2A 30+ 607,222 29.4 28,513 1.1 228,034 6.8 Total 2,056,870 100.0 2,594,464 100.0 3,338,965 100.0 Table 7: Spatial Distribution of Population Change: Brasilia, Curitiba and Recife, 1991-2000 BRAS[LIA CURITIBA RECIFE DISTANCE CATEGORY (KM) POPULATION % OF TOTAL POPULATION % OF TOTAL POPULATION % OF TOTAL CHANGE CHANGE CHANGE CHANGE CHANGE CHANGE 0-5 1,514 0.3 14,405 2.6 8,520 2.0 5.1-10 33,817 7.5 87,966 16.2 77,674 18.4 10.1-15 26,629 5.9 153,214 28.2 128,818 30.5 15.1-20 120,061 26.6 122,023 22.4 116,325 27.6 20.1-25 44,005 9.8 63, 135 11.6 31,677 7.5 25.1-30 126,885 28.1 87, 195 16.0 18,419 4.4 30+ 98,286 21.8 15,733 2.9 40,349 9.6 Total 451,197 100.0 543,672 100.0 421, 782 100.0 Comparison of the spatial distribution of 1991 and 2000 population and the change in population between 1991 and 2000 reveals several interesting results. The first and most dramatic finding is that Brasilia's population is distributed quite differently than Curitiba's and Recife's - most of its population is concentrated far from the city center. In 1991, over half (53.6%) of Brasilia's metropolitan population was located more than 25 kilometers from the city. By 2000, the percentage had declined somewhat, to 50%, but still remained distinctly different from the spatial patterns in the other two cities. The percentage of population located within 10 kilometers of Brasilia's center averaged about 8% for both 1991and2000. In sharp contrast, in 1991 nearly 70% of Curitiba's population resided within 10 kilometers of the city center. By 2000, Curitiba's population had begun to decentralize and 58.5% of the total metropolitan population was located within 10 kilometers of the center. Peripheral population in Curitiba was low in comparison to Brasilia - less than 6% in 1991 and less than 9% in 2000 of the total population residing more than 25 kilometers from the central city. In Recife, the patterns are similar. In 1991, over 48% of the population resided within 10 kilometers of the city center. In 2000, the portion was 44%. Recife's peripheral population was about the same as Curitiba's 78 Land and Urban Policies for Porvety Reduction and well below that of Brasilia. In 1991, 8.5% lived more than 25 kilometers from the city center. In 2000, the figure increased to 9.2%. The spatial distribution of population in the three cities between 1991 and 2000 largely reflected the baseline spatial structure of 1991. In Brasilia, about half of the population growth took place in areas more than 25 kilometers from the center. It is significant to note that approximately 27% of the population change took place in the distance band of 20.1-25 kilometers - reflecting the growth in the area northeast of the city center. This decentralized, sprawling pattern of population change suggests that planning restrictions and government ownership ofland introduces profound distortions into the urban land market. Since development is blocked in areas adjacent to the city center, residential growth is forced to the periphery. In Curitiba, population growth moved out beyond 10 kilometers from the city center. Between 1991 and 2000, nearly half of the increase took place in areas between 10 and 20 kilometers from the city. This suggests that Curitiba has been relatively successful in achieving compact development - channeling growth into areas that are contiguous to existing urban areas. In Recife, approximately 58% of the increase in population between 1991 and 2000 occurred between 10.1 and 20.1 kilometers from the city center. Like Curitiba, Recife's growth has been compact, moving out beyond the densely developed core. Tables 8, 9 and 10 provide breakdowns of developed urban land for the three cities for 1991, 1997, 2000 and change in urban developed land between 1991and1997/2000. In Brasilia's core of7,850 hectares (within 5 kilometers), less than 10% of the total urban land area is developed. In contrast, over 90% of the land in the core of Curitiba is developed. In Recife, the portion is nearly 40%. (About half of Recife's core, however, is in the ocean; thus, about 80% of its developable core is urbanized.) Over the 1991 to 1997/2000 period, very little additional land was urbanized. In Brasilia, net new urban development in the core - conversion of vacant land to urban uses - is effectively zero (1 hectare). In Curitiba, net urban development in the core increased by 14 hectares, and in Recife, was the greatest increase at 48 hectares. Table 8: Spatial Distribution of Urban Land Development: Brasflia, Curitiba and Recife, 1991 BRASILIA CURITIBA RECIFE DISTANCE URBAN LAND URBAN LAND URBAN LAND CATEGORY (KM) DEVELOPMENT % OF TOTAL DEVELOPMENT % OF TOTAL DEVELOPMENT % OF TOTAL (HA) (HA) (HA) 0-5 733 2.2 7,232 8.1 3,086 9.8 5.1-10 8,743 26.0 20,321 22.7 8,983 28.5 10.1:--15 5,707 17.0 19,260 21.5 6,854 21.7 15.1-20 6,929 20.6 21,594 24.1 5,057 16.0 20.H5 2,659 7.9 10,049 11.2 2,921 9.3 25.1-30 3,752 11.1 9,909 1.1 1,414 4.5 30+ 5, 144 15.3 1,294 1.4 3,244 10.3 Total 33,666 100.0 89,659. 100.0 31,559 100.0 Land and Urban Policies for Porvety Reduction 79 Table 9: Spatial Distribution of Urban Land Development: Brasilia, Curitiba and Recife, 1997 /2000 BRASILIA, 1997 CURITIBA, 2000 RECIFE, 1997 DISTANCE URBAN LAND URBAN LAND URBAN LAND CATEGORY (KM) DEVELOPMENT % OF TOTAL DEVELOPMENT % OF TOTAL DEVELOPMENT % OF TOTAL (HA) (HA) (HA) 0-5 734 1.4 7,246 6.6 3, 134 8.4 5.1-10 9,602 18.0 21,278 19.4 9,374 25.0 10.1-15 7, 151 13.4 23,325 21.3 8,275 22.1 15.1-20 10,926 20.5 28,861 26.3 7, 123 19.0 20.1-25 5,632 10.6 14,452 13.2 3.747 10.0 25.1-30 6,035 11.3 13,361 12.2 1,911 5.1 30+ 132,207 24.8 1, 106 1.0 3,855 10.3 Total 53,287 100.0 109,629 100.0 37,420 100.0 As far as urban land development beyond the core, Curitiba's and Recife's urban development is concentrated in the 10- to 25-kilometer bands. Between 1991 and 2000, 81 % of Curitiba's change in developed, urbanized land was located in this 10-25 kilometer band. In Recife, 73% was similarly located. In contrast, in Brasilia, less than 50% was located within 10 to 25 kilometers. In fact, approximately 53% of urban land development in Brasilia between 1991 and 1997 took place beyond 25 kilometers from the city center-suggesting that Brasllia is sprawling. Table 10: Spatial Distribution of Change in Urban Land Development: Brasflia, Curitiba and Recife, 1991-1997/2000 BRASILIA 1991 - 1997 CURITIBA 1991 - 2000 RECIFE 1991 - 1997 DISTANCE URBAN LAND URBAN LAND URBAN LAND % OF TOTAL % OF TOTAL % OF TOTAL CATEGORY (KM) DEVELOPMENT DEVELOPMENT DEVELOPMENT CHANGE CHANGE CHANGE CHANGE CHANGE CHANGE 0-5 .5 0.0 14 0.1 48 0.8 5.1-10 860 4.4 957 5.0 391 6.8 10.1-15 1,444 7.4 4,065 21.1 1.421 24.8 15.1-20 3,997 20.4 7, 133 37.1 1,942 33.9 20.1-25 2,973 15.2 4,403 22.9 827 14.4 25.1-30 2,283 11.6 2,836 14.8 497 8.7 30+ 8,063 41.1 -188 -1.0 611 10.7 Total 19,620 100.0 19,220 100.0 6,738 100.0 What are the implications of these alternative forms of urban land development in the three cities? There are three important issues that emerge from our comparison. First, cities that sprawl - such as Brasllia - consume more land per person than tho~e that develop compactly. Brasllia developed 19,620 hectares of land to accommodate 811,000 persons - 24 hectares per 1,000 additional persons. In contrast, Recife 80 Land and Urban Policies for Porvety Reduction developed 6,738 hectares of land to accommodate 422,000 additional persons - 16 hectares ofland per 1,000 persons. However, Curitiba developed 19,220 hectares ofland to accommodate 543,000 additional persons - 35 hectares of land per 1,000 persons suggesting that Curitiba experienced substantial low- density development. A second factor is the welfare implications of forcing population to travel greater distances to the center of the city. As Bertaud and Buckley have suggested for India, low-d~nsity urban sprawl introduces significant transportation costs on residents. A good comparative measure of compactness is the average per capita distance from the city center.. This is calculated as the weighted average distance of each population in each zone. In 2001, the average per capita distance for Brasilia was 24.3 kilometers; for Curitiba it was 11.2 kilometers; and for Recife it was 13.1 kilometers. In all cases, the average per capita distance to the city center increased between 1991 and 2001. In 1991, Brasllia's average was 22.5 kilometers, Curitiba's was 9.75 kilometers, and Recife's was 12.62 kilometers. In a recent paper, Bertaud and Bruckner illustrated that cities with restrictive development controls take up more space and have lower consumer welfare due to increased commuting costs. Given the fact that distances are approximately twice as great in Brasilia than they are in Curitiba or Recife, there is clearly a compelling case for assessing the welfare implications of the capital's dispersed spatial structure. The third impact is that more compact development economizes on urban infrastructure costs, whereas low- density sprawling development typically requires higher infrastructure costs per capita. Tables 11, 12 and 13 and Figure 2 provide tabulations of population density by distance from the city center for the three cities for 1991 and 2000. There are sharp density contrasts among Brasilia, Curitiba, and Recife. In the areas within 10 kilometers of the city center, densities in Curitiba and Recife are five to ten times greater than in Brasilia. Densities on the periphery of Brasllia are five to ten times higher than for Curitiba, and about twice as high as Recife. In the case of Curitiba, there is evidence of significant very low-density suburban development in the areas beyond 10 kilometers-despite its success with the development ofhigh- density development corridors. Table 11: Spatial Structure of Population Density: Brasflia, Curitiba and Recife, 1991 BRASILIA CURITIBA RECIFE DISTANCE CATEGORY (KM) POPULATION DENSITY POPULATION/ POPULATION DENSITY POPULATION/ POPULATION DENSITY POPULATION/ URBANIZED LAND (HECTARES) URBANIZED LAND (HECTARES) URBANIZED LAND (HECTARES) 0-5 6.2 64.5 108.8 5.1-10 13.5 47.4 119.0 10.1-15 20.0 14.0 108.0 15.1-20 30.9 8.1 73.6 20.1-25 103.5 5.7 51.9 25.1-30 95.2 10.7 42.5 30+ 48.4 9.9 57.9 Total 39.6 22.9 92.4 Land and Urban Policies for' Porvety Reduction 81 Table 12: Spatial Structure of Population Density: Brasflia, Curitiba and Recife, 2000 BRASILIA CURITIBA RECIFE DISTANCE POPULATION DENSITY POPULATION DENSITY POPULATION DENSITY CATEGORY POPULATION/URBANIZED LAND POPULATION/URBANIZED LAND POPULATION/URBANIZED LAND (HECTARES) (HECTARES) (HECTARES) 0-5 8.2 66.4 109.8 5.1-10 15.9 49.4 122.4 10.1-15 19.7 18.1 105.0 15.1-20 30.6 10.3 68.6 20.1-25 56.7 8.4 48.9 25.1-30 82.4 14.5 41.1 30+ 46.0 25.8 59.2 Total 38.6 23.7 89.2 Table 13: Change in Spatial Structure of Population Density: Brasflia, Curitiba and Recife, 1991-2000 BRASILIA CURITIBA RECIFE CHANGE IN POPULATION DENSITY CHANGE IN POPULATION DENSITY CHANGE IN POPULATION DENSITY DISTANCE CATEGORY POPULATION/URBANIZED LAND POPULATION/URBANIZED LAND POPULATION/URBANIZED LAND (HECTARES) (HECTARES) (HECTARES) 0-5 2.0 1.9 1.0 5.1-10 2.4 2.0 3.4 10.1-15 -0.3 4.1 -3.0 15.1-20 -0.3 2.2 -5.0 20.1-25 46.8 2.7 -3.0 25.1-30 -12.8 3.8 -1.4 30+ -2.4 15.9 1.3 Total -1.0 0.8 -3.2 Figure 2: Change in Spatial Structure of Population Density: Brasilia, Curitiba and Recife, 1991-2000 45 35 • Brasflia >- 25 +' 'Vi m Curitiba c Q) o Recife Cl 15 [ c 0 :.;:; ~ :J c. 5 0 a. ~..n fl LJ liiil II \J.- • • -5 -15 05 5.1-10 10.1-15 15.1-20 20.1-25 25.1-30 30+ Distance (Km) 82 Land and Urban Policies for Porvety Reduction Tables 14 - 19 and Figure 3 provide tabulations of formal and informal housing stock for the three cities for 1991, 2000, and the change between 1991 and 2000. All three cities substantially increased their housing stocks - from 300,000 to over 600,000 units during the 1990s. In Curitiba, most of the housing is located within 15 kilometers of the city- 86% in 1991 and 81 % in 2000. The situation in Recife is similar with 66% in 2000. In contrast, Brasilia's formal housing stock is predominantly located between 15 and 30 kilometers from the city center - 65% in 1991 and 66% in 2000. Less than 20% of the city's housing stock is located within 10 kilometers of the city center. The spatial patterns of informal housing are somewhat different from formal housing. Informal housing tends to be more concentrated near the centers of the metropolitan areas. In the cases of Curitiba and Recife, 94% and 89%, respectively, of the informal housing stock in 1991 was located within 15 kilometers of the center. By 2000, the percentage within 15 kilometers in both Curitiba and Recife slightly declined to 92% and 85%, respectively. In Brasilia, informal housing is effectively shunted to the periphery. In 1991, 61 % of informal housing was located more than 15 kilometers from the city center. In 2000, the corresponding figure was 68%. Table 14: Spatial Distribution of Formal Housing Stock: Brasilia and Curitiba, 1991 DISTANCE BRASILIA CURITIBA CATEGORY HOUSING HOUSING (KM) UNITS % OF TOTAL UNITS % OF TOTAL 0-5 1055 0.3 109,501 21.9 5.1-10 29,903 8.5 229,650 46.0 10.1-15 31,662 9.0 109,654 22.0 15.1-20 67, 194 19.1 26,014 5.2 20.1-25 68,953 19.6 8,557 1.7 25.1-30 92,878 26.4 14,473 2.9 30+ 60, 158 17.1 1,213 0.2 Total 351,803 100.0 499,062 100.0 Table 15: Spatial Distribution of Formal Housing Stock:Brasflia, Curitiba and Recife, 2000 DISTANCE BRASILIA CURITIBA RECIFE CATEGORY HOUSING HOUSING HOUSING (KM) % OF TOTAL % OF TOTAL % OF TOTAL UNITS UNITS UNITS 0-5 2, 101 0.4 137,618 20.1 67,343 10.1 5.1-10 48,855 9.3 259,085 37.8 201, 165 30.1 10.1-15 38,874 7.4 172,697 25.2 172, 173 25.8 15.1-20 111,368 21.2 51,662 7.5 121,535 18.2 20.1-25 85,627 16.3 37,258 5.4 38,919 5.8 25.1-30 148,666 28.3 26,026 3.8 18,846 2.8 30+ 89,830 17.1 545 0.1 47,837 7.2 Total 525,321 100.0 684,891 100.0 667,818 100.0 Land and Urban Policies for Porvety Reduction 83 Table 16: Spatial Distribution of Formal Housing Stock Change: Brasilia and Curitiba, 1991-2000 DISTANCE BRASILIA CURITIBA CATEGORY HOUSING HOUSING (KM) % OF TOTAL % OF TOTAL STOCK STOCK CHANGE CHANGE CHANGE CHANGE 0-5 1,215 0.7 28, 117 15.1 5.1-10 19,260 11.1 29,435 15.8 10.1-15 5,726 3.3 63,043 33.9 15.1-20 46,329 26.7 25,648 13.8 20.1-25 13,881 8.0 28,701 15.4 25.1-30 57,.435 33.1 11,553 6.2 30+ 29,672 17.1 -668 -0.2 Total 173,518 100.0 185,829 100.0 Table 17: Spatial Distribution of Informal Housing Stock: Brasilia and Curitiba, 1991 DISTANCE BRASILIA CURITIBA CATEGORY HOUSING HOUSING (KM) UNITS % OF TOTAL UNITS % OF TOTAL 0-5 0 0.0 3,586 10.5 5.1-10 169 2.9 16,998 49.8 10.1-15 2, 136 36.6 10,891 31.9 15.1-20 2,352 40.3 1,913 5.6 20.1-25 444 7.6 640 1.9 25.1-30 70 1.2 82 0.2 30+ 665 11.4 0 0 Total 5,836 100.0 34, 110 100.0 Table 18: Spatial Distribution of Informal Housing Stock: Brasilia, Curitiba and Recife, 2000 DISTANCE BRASILIA CURITIBA RECIFE CATEGORY HOUSING HOUSING HOUSING (KM) % OF TOTAL % OF TOTAL % OF TOTAL UNITS UNITS UNITS 0-5 0 0.0 3,893 8.1 29, 166 12.4 5.1-10 302 0.7 21, 192 44.2 111,594 47.5 10.1-15 13,414 31.1 18,543 38.7 57,898 24.7 15.1-20 12,940 30.0 2,526 5.3 17,911 7.6 20.1-25 8,066 18.7 1,315 2.7 6,548 2.8 25.1-30 0 0.0 467 1.0 2,240 1.0 30+ 8,411 19.5 0 0 9,364 4.0 Total 43, 132 100.0 47,936 100.0 234,721 100.0 84 Land and Urban Policies for Porvety Reduction Figure 3: Spatial Distribution of Informal Housing Stock:Brasflia, Curitiba and Recife, 2000 55.0 50.0 45.0 40.0 35.0 !f: 30.0 -----·-·--..·--·-·-_.. ___._ _,,._._ ______ .. .. _, , I 1 QJ 25.0 1-----1 ·:':1 c.. ..., c"' QJ ~ °' QJ 20.0 ,, ··-·- • Brasilia El Curitiba 15.0 D Recife - 10.0 r- SJ - 1,X ~I ,- 5.0 0.0 -·>'": 1+; bD, .,---, 0-5 5.1-10 10.1-15 15.1-20 20.1-25 25.1-30 30+ Distance (Km) Table 19: Spatial Distribution of Informal Housing Stock Change: Brasflia and Curitiba, 1991-2000 DISTANCE BRASILIA CURITIBA CATEGORY HOUSING STOCK % OF TOTAL HOUSING STOCK % OF TOTAL (KM) CHANGE CHANGE CHANGE CHANGE 0-5 0 0.0 307 2.2 5.1-10 133 -1.4 4, 194 30.3 10.1-15 11,278 26.0 7,652 55.3 15.1-20 10,587 20.3 613 4.4 20.1-25 7,622 29.2 675 4.9 25.1-30 -70 -1.1 30+ 7,746 27.0 385* 2.8* Total 37,296 100.0 13,826 100.0 * 25+ Land and Urban Policies for Porvety Reduction 85 The Effects of Location, Titling, Infrastructure and Plot Size on Residential Land Prices in the Three Cities Land value data from the three cities were gathered through a systematic survey of real estate brokers. Price data were gathered for various types of residential plots in each geographic zone of each city. Price data are therefore available by distance from the city center and according to whether plots are legally titled, have access to infrastructure (electric, water, paved roads), and whether they are under or over 500 square meters in size. Data were collected for two time periods between 2000 and 2003. All price data presented in the report are in 2003 constant prices. Over the 2000-2003 period, the IPCA (Indice Nacional de Pre~os) increased by 32.9% and the IPCA has been used to adjust prices upwards to 2003 terms. The section starts by presenting overall descriptive statistics of residential land values. It then proceeds to report on the results of three multivariate regression models that seek to gauge the independent effects of distance, title, infrastructure, and plot size. Table 20 presents mean plot prices for the three cities. Interestingly, current plot prices (unadjusted for inflation) did not increase as rapidly as the IPCA, and therefore are higher in the earlier years-for Brasi'.lia, 209 reais per square meter in 2001 and 142 reais in 2003; for Curitiba, 67 reais per square meter in 2000 and 66 reais in 2002; and for Recife, 74 reais per square meter in 2001 and 71 reais in 2003. This suggests that real plot prices have not kept pace with inflation. It is also noteworthy that residential land prices are considerably higher in Brasilia than in Recife and Curitiba, averaging 142, 71, and 66 reais, respectively, in 2003. While there are a myriad of factors shaping residential land prices, high per capita and household incomes in the capital probably explain much of the difference-higher incomes mean higher ability to pay for housing, driving up land prices. It may also be the case that strict land use planning and development controls in Brasilia limited the supply of land for residential development, particularly in the more centrally located areas, and resulted in higher land prices. Table 20: Mean Plot Prices by Infrastructure Service and Title in 2000, 2001, 2002 and 2003 for Brasflia, Curitiba and Recife in Constant 2003 Values (Reais per Square Meter) DISTANCE BRASILIA CURITIBA RECIFE CATEGORY (KM) 2001 2003 2000 2002 2001 2003 All plots 209 142 67 66 74 71 Infrastructure With 164' 139' ·109 108 102 97 Without 157' 47' 38 38 44 42 Title With 223 147 68 68 81 78 Without 193 122 66 66 67 64 Infrastructure With 213 144 73 73 77 73 and Title without * * 29 29 18 17 Plot Size < 500m2 184 153 72 71 75 71 >500m2 266 99 57 57 73 70 'Means exclude plots located within 1O kilometers of the city center. * Sample size less than 30 Deflators: 2000=1.329; 2001=1.244; 2002=1.147 and 2003=1.000. 86 Land and Urban Policies for Porvety Reduction Provision of infrastructure has a clear and positive impact on residential plot prices in the three cities. As illustrated in Table 20 the mean 2002/2003 value of plots with infrastructure (measured by the presence of paved roads) ranges from 139 reais in Brasilia, to 108 reais in Curitiba, and 97 reais in Recife. These means are all greater than corresponding prices for plots without infrastructure: in Brasilia, 47 reais; in Curitiba, 38 reais; and in Recife, 42 reais. These patterns reflect the positive impact that infrastructure provision has on land values. Below, we present a more rigorous analytical examination of the effects of infrastructure on land prices. Although to a lesser extent, the provision of title of property ownership also positively affects residential land prices. As illustrated in Table 20 the mean 2002/2003 value of plots with title ranges from 147 reais in Brasilia, to 68 reais in Curitiba, and 78 reais in Recife. These means are all greater than corresponding prices for plots without title: in Brasilia, 122 reais; in Curitiba, 66 reais; and in Recife, 64 reais. We have also found that the existence of both infrastructure and title positively affects prices. In Table 20 the mean 2002/2003 value differential for plots with both infrastructure and tide and without is 73 reais versus 29 reais for Curitiba and 73 reais versus 17 reais for Recife. We also found that plot size affects per-meter prices of plots, although the impact is variable across the three cities (see Table 20). In the case of Brasilia, large plots have higher prices per square meter - 266 reais as compared to 184 reais for plots under 500 square meters. Since it is normally the case that smaller plots have higher prices per square meter, the results in Brasilia may reflect the fact that there is a scarcity of large plots in the metropolitan area. In both Curitiba and Recife, per-square-meter plot prices are higher for small plots than for large plots - 72 reais versus 57 reais for Curitiba and 71 reais versus 70 reais for Recife. The above results are highly general since they do not incorporate the effects of location into the calculations of means. Table 21 provides tabulations of mean plot prices per square meter based on distance from the central city. Plot prices in Brasilia display the same unique patterns as for population density and housing. Plot prices in 2003 incr:ease as distance from the center increases, up to 10 kilometers, and gradually decline out to 30 kilometers. Beyond 30 kilometers, plot prices are much lower (but still more than double comparable prices in Curitiba and Recife). This distinct pattern ofland prices is the result of strict land use planning controls in Brasilia, limitations on housing in the core, and strict exclusion of informal housing within the federal district. Beyond the federal district, informal housing is more common and there is an active market for unauthorized houses and condominiums. Plot prices in Curitiba and Recife display more conventional patterns. Prices are highest at the center of the city and then decline consistently with increasing distances. In 2002, prices in Curitiba were 180 reais per square meter; beyond 10 kilometers, prices range from 44 reais to 11 reais per square meter, depending on location. Interestingly, prices beyond 30 kilometers are higher than in the 25-30 kilometer band, 17 reais versus 11. Further investigation is needed to ascertain what factors cause this up-tick in prices. In Recife, plot prices are highest in the city center at 169 reais and decline steadily with increasing distance. At the periphery, plot prices average 28 reais, considerably higher than Curitiba, but well below comparable levels in Brasilia. . Comparing plot prices over time (2000-2002 and 2001-2003), suggests that prices in both Curitiba and Recife have been fairly c~nstant in real, inflation adjusted terms. In Brasilia, real inflation adjusted prices appear to have declined in suburban areas, while increasing in the core (0- 5 kilometers). Land and Urban Policies for Porvety Reduction 87 Table 21 presents tabulations of residential land prices by distance from the city center. They reveal a striking difference between the highly planned city of Brasilia and Curitiba and Recife. In the cases of Curitiba and Recife, residential land prices systematically decline from the city center. In Curitiba, 2002 prices of plots located within five kilometers of the city center average 180 reais. Farther out, from 5-10 kilometers, the mean price falls to 78 reais. This pattern continues all the way out to the 25-30 kilometer band, where prices fall to 11 reais. However, beyond 30 kilometers, prices pick up a bit to 17 reais. In Recife, plot prices decline from 169 reais to 28 reais at the edge. Table 21: Mean Plot Prices by Distance from City Center in 2000, 2001, 2002, and 2003 for Brasilia, Curitiba and Recife in Constant 2003 Values (Reais per Square Meter) DISTANCE BRASILIA CURITIBA RECIFE CATEGORY (KM) 2001 2003 2000 2002 2001 2003 0-5 222 317 185 180 173 169 5.1-10 802 512 78 78 96 91 10.1-15 156 203 42 44 45 42 15.1-20 176 176 18 18 27 26 20.1-25 124 145 15 16 34 31 25.1-30 168 121 12 11 38 40 30+ 200 64 19 17 27 28 Total 233 142 67 66 74 71 Deflators: 2000=1.329; 2001=1.244; 2002=1.147 and 2003=1.000. In Brasilia, land prices increase from the center out to 10 kilometers. From 10 to 30 kilometers, prices remain very high (particularly in comparison to the other two cities). This pattern appears to reflect the strict land use development regulations that exist in Brasilia, with opportunities for housing restricted to limited areas in and around the center, and most residential development located 10 to 20 kilometers from the center. This pattern in prices reflects the population density aspects discussed above in a previous section. The tabulations of mean plot prices according to plot characteristics and location indicate that prices are strongly affected by these factors. In the remaining portion of this section, we attempt to isolate the effects of each of these factors by building two multivariate regression models to predict residential plot price-one for 2002-2003 price data and one for 2000-2001 price data. In developing the models, we took an exploratory approach, utilizing two functional forms (linear and log-linear) and a step-wise process for determining which independent variable to include in the models. In the case of the linear model, the dependent variable was constant per-square-meter plot price. Independent variables include distance from city center and a range of dummy variables to indicate the presence of a range of plot characteristics-provision of infrastructure (electric, water, paved roads), availability of title, and plot size (over or under 500 square meters). In order to test for potential interaction effects, we also included a variable that captured the presence of both title and infrastructure. In the case of the log-linear model, the dependent variable was the log (base e) of constant per- square meter plot price. The log-linear model used the same independent variables. 88 Land and Urban Policies for Porvety Reduction The step-wise process iteratively adds independent variables to the regression model in an attempt to build the most robust model. It results in various model specifications, depending on the explanatory power of added variables. Model runs indicated that the log-linear specification was the most robust. Tables 22 and 23 present the results of the log-linear models for Brasilia, Curitiba and Recife for 2002-2003 and 2000-2001. In Table 22, the step-wise modeling results in three distinct models for each city. In the. case of Brasilia, the best model incorporated distance, pavement dummy and plot size. It excluded electric, water, tide, and pavement-tide. Overall, the model is highly predictive, with an adjusted R2 of 0.585. All of the independent variables are significant at the .000 confidence level and have the expected signs (constant is positive, distance is negative, pavement is positive, and plot size is negative). The 2002 model for Curitiba is also very significant. It has an R2 of .656. The Curitiba model includes constant, distance, pavement, pavement-tide and plot size. All of the independent variables are significant at the .000 confidence level and have the expected signs. The Recife model has the lowest R2 of the three cities-.394. It includes constant, distance, pavement, and tide. All of the independent variables are significant at the .000 confidence level and have the expected signs. Table 22: Stepwise Regression Results: Brasilia, Curitiba and Recife, 2002 and 2003, Dependent Variable: Log (base e) of Constant Reais per Square Meter BRASILIA 2003 CURITIBA 2002 RECIFE 2003 BETA SIGNIFICANCE BETA SIGNIFICANCE BETA SIGNIFICANCE CONSTANT 6.055 .000 4.469 .000 3.968 .000 (27.873) (76. 191) (108.869) DISTANCE TO CBD -.089 .000 -.117 .000 -.047 .000 (-13.737) (-49.931) (-28.910) PAVEMENT DUMMY 1.027 .000 .748 .000 .639 .000 (7.373) (19.676) (20.710) TITLE DUMMY .194 .000 (6.440) PAVEMENT AND TITLE DUMMY .391 .000 (7.215) PLOT SIZE DUMMY -.993 .000 -.419 .000 (-6.313) (-11.295) ADJUSTED R2 .585 df =175 .656 df= 1921 .394 df= 1921 T statistics are in parentheses. I ' I I Land and Urban Policies for Porvety Reduction 89 I Table 23 presents model results for 2000-2001 years. There are no results for Brasilia. The results for Curitiba and Recife, however, were robust and are generally similar to the results for 2002-2003 presented in Table 22. I Tables 24 and 25 interpret the results of the five models. The constant values are presented in the first row of I both tables. The constant value is the estimated value of the plot located at the center of the city, with no paved road, no title and small plot size. For the 2002-2003 years, the constants range from 426 reais for l Brasllia, 87 reais for Curitiba, and 53 reais for Recife. For 2000-2001 (Table 25), the constant values are 86 reais for Curitiba and 56 reais for Recife. Table 23: Stepwise Regression Results: Brasflia, Curitiba and Recife, 2000 and 2001, Dependent Variable: Log (base e) of Constant (Reais per Square Meter) BRASiLIA 2001 CURITIBA 2000 RECIFE 2001 BETA SIGNIFICANCE BETA SIGNIFICANCE BETA SIGNIFICANCE CONSTANT 4.451 .000 4.024 .000 (75.473) (109.862) DISTANCE TO CBD -.116 .000 -.048 .000 (-49.243) (-29.356) PAVEMENT DUMMY .760 .000 .636 .000 (19.883) (20.491) TITLE DUMMY .198 .000 (6.515) PAVEMENT AND .396 .000 TITLE DUMMY (7.269) PLOT SIZE DUMMY -.439 .000 (-11.766) ADJUSTED R, .653 df=1921 .396 df=2500 T statistics are in parentheses. 90 Land and Urban Policies for Porvety Reduction Table 24. Interpreted Regression Results: Brasflia, Curitiba and Recife, 2002 and 2003, (from Table 22) Bold figures are estimates of dependent variable (plot price per square meter [Reais]) 1 I BRASILIA 2003 CURITIBA 2002 RECIFE 2003 Constant value, no paved road, no title and small plot (reais per square 426 87 53 meter) Value adjustment for having paved 2.79=> 1,189 2.11=>184 1.89 => 100 road (factor and reais per square meter) Value adjustment for having title 1.21=>64 (factor and reais per square meter) Value adjustment for having both 1.47 => 128 paved road and title (factor and reais per square meter) Value adjustment for having large .370 => 158 .658 => 57 plot (factor and reais per square meter) -.089 intercept -.117 intercept -.047 intercept value Distance value adjustment per value at 10 kilometers value at 10 kilometers at 10 kilometers kilometer from city center (factor and 175 27 33 reais per square meter) The second row of Tables 24 and 25 presents estimates of the effect of having a paved road on plot prices. There are two numbers in each cell. The first number is the shift effect (adjustment value) of having a paved road. For example, in the case ofBrasflia in 2003, the constant value is multiplied by 2.79 (a 179% increase) to estimate the adjusted price of a plot located at the city center, with a paved road, no title and small plot size. The estimated value is 1, 189 reais. The effects of pavement (which should be thought of as a proxy for infrastructure) are very strong in all three cities. Presence of infrastructure adds a land price premium of 179% in Brasilia, 111 % in Curitiba, and 89% in Recife. Land and Urban Policies for Porvety Reduction 91 Table 25: Interpreted Regression Results:Brasflia, Curitiba and Recife, 2000 and 2001, (from Table 23) Bold figures are estimates of dependent variable (plot price per square meter [Reais]) BRASILIA 2001 CURITIBA 2000 RECIFE 2001 Constant value, no paved road, no title and small plot (reais per square 86 56 meter) Value adjustment for having paved 2.14 => 184 1.89=> 106 road (factor and reais per square meter) Value adjustment for having title (factor and reais per square meter) 1.22 => 62 Value adjustment for having both paved road and title (factor and reais 1.49 => 128 per square meter) Value adjustment for large plot .645 => 55 (factor and reais per square meter) Distance value adjustment per -.116 intercept -.048 intercept value kilometer from city center (factor and value at 10 kilometers 3t 10 kilometers reais per square meter) 27 53 The third row of Tables 24 and 25 presents estimates of the effect of title on land prices. This effect shows up only in Recife and indicates that title adds about 20% to the price of a plot. However, if we combine the effects of pavement and title, effects show up in Curitiba. As row four of Tables 24 and 25 indicate, the value adjustment for having both infrastructure and title increases plot prices by 47%-49%. It is interesting to note that, unlike pavement, title does not generate as consistent and large effects. While this result requires further exploration, it may be the case that Brasilia's planning and regulatory system overwhelms the effects of title. Virtually all plots in the federal district have title, and the presence or absence of title is only relevant on the fringes of Brasilia's metropolitan area. In Curitiba, title on its own does not generate statistically significant effects. Only when combined with infrastructure does the effect surface. Here it may be the case that titled but unserviced plots have prices that are similar to untitled and unserviced plots. The fifth row of Tables 24 and 25 provides estimates of the effects of plot size on plot price per square meter. In Brasilia for 2003 and Curitiba for both 2000 and 2002 and in Recife for 2001 and 2003, the price oflarge plots per square meter is well below the per-square-meter price of smaller plots. This seems to reflect market experience elsewhere. 92 Land and Urban Policies for Porvety Reduction Finally, row six ofTables 24 and 25 provides estimates of the effect oflocation (measured in terms of distance from the city center) on plot prices. These adjustment factors, referred to as price gradients, estimate the percentage change in plot prices relative to increases in distance. For example, in the case of Brasilia in 2003, for each one kilometer increase in distance from the city center, the price of a plot decreases by 8.9%. At a distance of 10 kilometers from the city, the constant price is reduced to 175 reais (versus 426 reais at the city center).At 10 kilometers from the center, the constant is worth only 41 % of its city center value. In Curitiba, the gradient is -.117, and at 10 kilometers from the center, the constant is reduced to 27 reais (versus 87 reais) -it is worth only 31 % of its city center value. In the case of Recife, the gradient is -.047. At 10 kilometers, the constant is worth 33 reais - 62% of its city center value. Interestingly, the slope gradient for Curitiba is high in absolute terms (-.117), indicating that distance drives down prices more per kilometer than in either Brasllia (-.089) or Recife (-.047). This seems counter-intuitive given Curitiba's reputation for an efficient mass transit system. The result may be more of a reflection of the relatively high wages in Curitiba and therefore the higher opportunity cost of travel time. Recife's low price gradient is most likely due to its lower incomes and lower opportunity costs of travel. Conclusions This report has presented the results of land market assessments in three Brazilian cities. There are several overarching conclusions that can be drawn from the effort. First, it is feasible to carry out such assessments. Second, they result in the compilation of socio-economic, land use and land price information that is useful for gauging the effectiveness of urban planning, infrastructure provision and land titling. Third, the results indicate that urban land market dynamics in less regulated cities (Curitiba and Recife) perform well and reflect patterns and trends found in other cities around the world. Spatial patterns of urban development dramatically vary between the highly planned Brasllia and the more market-driven cities of Curitiba and Recife. Average distance per capita in Brasllia is more than double the levels of Curitiba and Recife. Data on formal housing stock patterns indicate that housing is abundant in the core areas of Curitiba and Recife-over half of Curitiba's stock is located within 10 kilometers of the city center, and in Recife, 40% is located within 10 kilometers. In contrast, less than 10% of Brasllia's formal housing stock is located within 10 kilometers of the center. Prices of residential land in suburban areas of Curitiba and Recife are in the 30-40 reais per square meter range. For plots of 400 square meters, this works out to between 12,000-16,000 reais (US$4,000- US$5,300). In the case of Brasilia, significant land market distortions were identified. Population is forced to commute longer distances and land prices are about 5 times higher in suburban areas than in Curitiba and Recife. Plots in suburban areas of Brasllia range from 150-200 reais per square meter. For 400 square meter plots, prices average 60,000-80,000 reais (US$20,000 to US$26,700). With respect to infrastructure provision and its effects on land prices, the results indicate that infrastructure investment have significant positive effects on land values. The results in the three cities indicate that infrastructure provision can increase land prices by 89%-179%. This suggests that there is ample scope for financing infrastructure provision through property· taxation, land value capture or other fiscal mechanisms. With respect to provision of title, the evidence is less compelling. In the case of Recife, the analysis consistently identified statistically significant positive effects generated by titling. There, the provision of infrastructure Land and Urban Policies for Porvety Reduction 93 increased land prices by approximately 20%. In the case of Curitiba, the joint provision of infrastructure and title increased prices by nearly 50%. Again, this suggests that there is scope for financing titling projects through some form of property taxation or value capture. References Bertaud, Alain. 2001. Metropolis: A Measure of the Spatial Organization of 7 Large Cities. http://alain- bertaud.com. Bertaud, Alain, and Jan K. Brueckner. 2004 (April). Analyzing Building Height Restrictions: Predicted Impacts, Welfare Costs, and a Case Study of Bangalore. World Bank Policy Research Paper 3290. Bertaud, Alain, and Bertrand Renaud. Socialist Cities Without Land Markets. Journal of Urban Economics 41:137-151. Braga, Andrea da Costa, and Fernando A. R. Falcao. 1997. Guia de Urbanismo, Arte eArquitetura de Brasilia. Brasilia: Fundac;:ao Athas Bulcao. Brinkhoff, Thomas. 2004. http://www.citypopulation.de/index.html, accessed June 2004. Dowall, Davjd E. 2003. Concept Paper on Brazil Land Market Assessment. Washington: World Bank. Dowall, David E. 1995. The Urban Land Market Assessment: A New Tool for Urban Management. Wa- shington and Nairobi: World Bank and UN CHS. Fernandes, Edesio. 2001 (December). New Statute Aims to Make Brazilian Cities More Inclusive. Habitat Debate 7(4):19. Fundac;:ao de Desenvolvimento Municipal (FIDEM). 1999. Estudo sabre Caracterizaqao da Pobreza Urbana na RMR- PROMETR6POLE. Recife, Brazil: FIDEM. - - - . 2003. Analise do mercado de solo urbano na regiao metropolitana do Recife. Recife, Brazil: FIDEM. GDF-Departamento de Urbanismo. 1992. Plano de Ordenamento Territorial do Distrito Federal- Secreta- ria de Obras e Servic;:os Publicos - Brasilia. Grupo de Trabalho sabre Habitaqao para Formular Polltica Nacional de Desenvolvimento Urbano para o Brasil. 1996. Relat6rio. Brasilia. Mimeo. Grupo de Trabalho Caixa Economia Federal, IPED, FINDIEC-UnB, World Bank. 2002. Polfrica Habitaciona1 para o Brasil. Brasilia: Mimeo. Instituto Brasileiro de Geografia e Estatistica (IBGE). 1991, 2000. Censos Demograficos 1991, 2000. Brasilia: IBGE. - - - . 2001. Perfil dos Municipios Brasileiros. Mimeo. Brasilia: IGBE. Instituto de Planejamento Territorial e Urbano do Distrito Federal. 1997. Plano Diretor de Ordenamento Territorial do Distrito Federal - PDOT. Brasilia: IPDF. 94 Land and Urban Policies for Porvety Reduction IPEA/IBGE/ NESUR-IE-UNICAMP. 1998. Estudo Caracterizai;:ao e Tendencias da Rede Urbana do Brasil. IPEA, USP, UFPA, UFPE. 200la. Configurai;:ao Atual e Tendencias da Rede Urbana. Serie Caracteriazai;:ao e Tendencias de Rede Urbana do Brasil. Volume 1. Brasilia: IPEA. - - - . 200lb. Gestao de Uso do Solo e disfuni;:oes do Crescimento Urbano: Instrumentos de Planejamen- to e Gestao Urbana: Belem, Natal e Recife. Serie Caracteriazai;:ao e Tendencias de Rede Urbana do Brasil. Volume 2. Brasilia: IPEA. - - - . 200lc. Gestao de Uso do Solo e disfuni;:oes do Crescimento Urbano: Instrumentos de Planejamento e Gestao Urbana: Brasllia e Rio de Janeiro. Serie Caracteriazai;:ao e Tendencias de Rede Urbana do Brasil. Volume 3. Brasllia: IPEA. - - - . 200ld. Gestao de Uso do Solo e disfuni;:oes do Crescimento Urbano: Instrumentos de Planejamen- to e Gestao Urbana: Curitiba. Serie Caracteriazai;:ao e Tendencias de Rede Urbana do Brasil. Volume 5. Brasllia: IPEA. Mills, Edwin, and Byong-Nak Song. 1972. Urbanization and Urban Problems. Cambridge: Harvard University Press. Motta, Diana. 2003. An Empirical Study of Land Markets and Land Policy in Three Metropolitan Areas: Curitiba, Brasflia, Recife. Materials for presentation at The World Bank Urban Research Symposium. Topic IV: Land Access and Land Use. Washington, D.C. December 15-17, 2003. Prefeitura Municipal de Valparafso de Goias (PMVG); Secretaria de Planejamento e Desenvolvimento de Goias; Ambiente Urbano Planejamento e Projetos SC. Ltda. Plano diretor do munidpio de Valparaiso de Goias. Produto 01. Caracterizai;:ao do Munidpio. 2003. Prefeitura Municipal de Santo Antonio do Descoberto - Goias. 1997. Plano diretor urbano - 1997 Mem6ria. Secretaria de Planejamento e Desenvolvimento de Goias I Fernando Teixeira e Associados. 2000. Plano dire- tor de Aguas Lindas de Goias. SEDU/Presidencia da Republica, Fundai;:ao Joao Pinheiro. 2001. Deficit Habitacional no Brasil 2000. Brasflia: SEDU/Presidencia da Republica, Fundai;:ao Joao Pinheiro. Secretaria de Estado de Desenvolvimento Urbano e Habitai;:ao (SEDUH)/ Subsecretaria de urbanismo e Preservai;:ao (SUDUR). 2003. Entorno do Distrito Federal. Volume I. Brasflia: SEDUH, SUDUR. SEPLAN-PCR. 1993. Indicador Dl - Uso da terra (Km2). Recife: URB-Recife/SEPLAN. Simmonds, Roger, and Gary Hack. 2000. Global City Regions: Their Emerging Forms. London: Spon. Souza, Flavia A.M. de. 2001 (November). The Future ofinformal Settlements: Lessons in the Legalization of Disputed Urban Land in Recife, Brazil. Geoforum 32(4):483-492. Subsecretaria de Polftica Urbana e Informai;:ao (SUPIN) and Secretaria de Estado de Desenvolvimento Urba- no e Habitai;:ao (SEDUH). 2003 (August). Relatorio final: analise do mercado de solo urbano no distrito federal e entorno imediato. Land and Urban Policies for Porvety Reduction 95 World Bank. 2002. Proposta para uma Nova Politica de Habitac;:ao para o Brasil. Mimeo. Washington: World Bank. World Bank/GTZ/FIDEM. 2002. Projeto Mercado lmobiliario Informal. - a Inclusao Social do Morador dos Loteamentos Clandestinos e lrregulares da RMR. World Bank/Cities Alliance/IPEA/FIDEM. 2002. Plano METR6POLE ESTRATEGICA. AGGLOMERATION AND URBAN PRODUCTIVITY: IMPLICATIONS FOR THE APPRAISAL OF TRANSPORT INVESTMENT 1 Daniel J Graham* Abstract This paper is concerned with the links between city size, productivity and infrastructure provision. The role of transport infrastructure in sustaining productivity is notoriously hard to isolate empirically due to the inter-dependent nature of the relationship, which creates problems of simultaneity bias. In this paper we show that transport investment can have an important influence on productivity by increasing the effective density of jobs within a given distance. We estimate elasticities of productivity with respect to city size for different industrial sectors of the economy using data on UK firms. The results show evidence of positive agglomeration externalities across a range of industries. Introduction This paper is concerned with the link between city size and productivity. It provides a quantitative assessment of this relationship for different sectors of the UK economy. The motivatio~ for exploring this theme is to identify if there might be any external benefits that arise from the provision of transport infrastructure that are not included in standard transport appraisals. Specifically, we investigate whether there is an association between the density of economic activity and levels of productivity. This theme is important in assessing the benefits of transport investment for two reasons. First, because ultimately transport investment is crucial in sustaining cities and supporting urban agglomerations and these in turn may provide external benefits to the economy. Second, because it is clear that a change in the level of transport infrastructure in any area will typically change the effective density of economic activity that is accessible to that area with associated implications for productivity and efficiency. Venables (2003) has shown that estimates of the elasticity of productivity with respect to city size can be used to shed light on the external benefits of transport improvements. He develops a computational model of an urban economy that links productivity to transport investment via effects on city size. His objective is to distinguish the real income changes that result from transport investment due to a productivity-city size effect, from those economic benefits that are captured in standard transport appraisal and which arise from resources saved in commuting and from an increase in urban output. An outline of Venables' model is given in figures 1 to 3. Figure 1 shows an urban equilibrium in which the size of the city is determined at point E, where the wage gap between city workers and non-city workers is entirely dissipated in the travel costs of the city worker who is most distant from the CBD. 1 This paper was given at a World Bank Urban Research Symposium held in April 2005. It presents initial results from research commissioned by the UK Department for Transport into the link between agglomeration and productivity. Results from latter stages of this research obtained using more extensive data, and details concerning the data sources and methodology, are given in Graham (2005, 2006a, forthcoming). * Centre for Transport Studies, Imperial College London, London, SW7 2AZ, UK. Email: d.j.graham@imperial.ac.uk. 98 Land and Urban Policies for Porvety Reduction Figure 2 shows that when a transport improvement is made commuting costs are shifted downwards and consequently the city expands to point E*. The increase in the output of the urban economy is area p + y. Note that since productivity is higher in the city by an amount equal to the height of the wage gap (WW), if workers transfer from outside the city to inside they are more productive. The total change in the resources used in commuting is y - a, which combined with the change in out put (~ + y), yields a benefit (real income gain) from the transport improvement of a+~- In figure 3 Venables considers the implications of the existence of a city size - productivity gradient. If, as the literature suggests, larger cities have higher productivity the productivity gap is now expressed as a concave curve that increases with city size. Equilibrium is found at the intersection of the commuting cost and wage gap curves. The fact that productivity is non-constant with respect to city size means that the real income gain from a transport improvement is a+~+(), where 8 measures the increase in productivity experienced by city workers and is akin to a measure of the elasticity of productivity with respect to city size. As Venables points out, the additional benefit 8 is the effect that would be missed by a standard transport appraisal. Estimates of 8 do exist, but they tend to be exclusively for manufacturing industries. The purpose of this paper is to quantify 8 for detailed sectors of the economy to show the extent to which the city size productivity effect is prevalent. Specifically, we estimate elasticities of productivity with respect to a measure of the effective density of jobs. We estimate the magnitude of these values for different industrial sectors. The paper is structured as follows. Section two discusses the literature on agglomeration productivity and transport investment. Sources of data used in the analysis are described in section three. The estimation model is next discussed in section 4. Estimation results are presented in section five. Conclusions are drawn in the final section. Agglomeration productivity and transport investment Agglomeration and productivity The tendency towards concentration or agglomeration is perhaps the most widely observed feature of the spatial organisation of economic activity. It can be discerned across the Globe at a variety of different geographical levels. Agglomeration is evident, for instance, in the existence and growth of cities, in the formation of industrial regions and districts, and in the clustering of like activities within the same neighbourhood of a town or city. The theory of agglomeration economies is based on the premise that the tendency towards spatial concentration is caused by the existence of positive externalities that are generated through close spatial proximity and that serve to raise the efficiency of firms. The externalities generated through agglomeration are traditionally categorised under three headings. Internal scale economies describe efficiency gains that occur as the overall scale of production is increased. With respect to agglomeration, the crucial assumption regarding internal scale economies is that they are internal at the plant level and therefore imply production at a single location rather than being spread across a number oflocations. Localization economies describe efficiency gains generated through the increased scale of a particular industry operating in close spatial proximity. Benefits are thought to be generated in three ways: through 'technological spillovers' between firms within the same industry, through the efficient provision of intermediate inputs to Land and Urban Policies for Porvety Reduction 99 firms in greater variety and at lower cost due to the growth of subsidiary trades, and through labour market pooling. Localization economies are intra-industry; they are external to firms but internal to the industry. Urbanization economies describe the productive advantages that accrue to firms through location in large population centres such as cities. Firms derive benefits from the scale of markets, from the proximity of market areas for inputs and outputs, and from good infrastructure and public service provision. These spatial external economies are cross-industry; they are external to the firm and the industry but internal to cities. The empirical literature on the link between agglomeration and productivity has been comprehensively reviewed in previous surveys by Rosenthal and Strange (2004), Eberts and McMillen (1999), Henderson (1988), Gerking (1994), and Moomaw (1983a). Here we briefly summarise some of the main results on the magnitude of estimated values of the effects of agglomerate economies on productivity Econometric studies of the effects of agglomeration on productivity have been conducted almost exclusively for manufacturing industries. Table 1 provides a summary of results from the literature relating to the effects of agglomeratioi;i' on productivity. We summarise here those studies that have produced an actual elasticity estimate of the effects of agglomeration rather than those that have detected agglomeration effects through the use of dummy variables or other limited variable methods. Table 1: Estimates of agglomeration economies from production function analyses. AUTHOR UNIT OF ANALYSIS DEPENDENT VARIABLE INDEPENDENT VARIABLE ELASTICITY Aaberg (1973) Swedish cities productivity city size (population) O.Q2 Shefer (1973) US MSAs productivity RTS at MSA aggregation 0.20 Sveikauskas (1975) US MSAs productivity city size (population) 0.06 Kawashima (1975) US MSAs productivity city size (population) 0.20 Fogarty and Garofalo (1978) US MSAs productivity city size (population) 0.10 Moomaw(1981) US MSAs productivity city size (population) 0.03 Moomaw (1985) US MSAs productivity city size (population) 0.07 Nakamura (1985) Japanese Cities productivity city size (population) 0.03a Tabuchi (1986) Japanese Cities productivity city size (population) 0.04 Lauri (1988) Greek Regions productivity city size (population) 0.05 Sveikauskas et al (1988) US MSAs productivity city size (population) 0.01b Nakamura (1985) Japanese Cities productivity industry size (employment) 0.05 Henderson (1986) Brazilian Cities productivity industry size (employment) 0.1 lc Henderson (1986) US MSAs productivity industry size (employment) 0.19d Henderson (2003) US MSAs plant output industry size (no. of plants) 0.03e Ciccone and Hall (1996) US States productivity employment density 0.06 Ciccone (2002) EU regions productivity employment density 0.05 Notes: a - mean value for 14 industries, b - mean value from 5 model specifications, c - mean value for ten industries, d - mean value for 9 industries, e - mean value for 4 model specifications. 100 Land and Urban Policies for Porvety Reduction With the exception of Shefer (1973) regressions use metropolitan population as a proxy for city size within a basic production function framework. The availability of capital stock data at the metropolitan level has exerted a big influence on the functional form used for estimation 1• Early studies were typically based on either the Constant Elasticity of Substitution (CES) function (e.g. Shefer 1973, Sveikauskas 1975, Moomaw 1981) or the Cobb-Douglas (CD) (e.g. Aaberg 1973, Kawashima 1975, Fogarty and Garofalo 1978). More recent estimates are based on flexible functional forms such as the translog (e.g. Nakamura 1985, Henderson 1986, Tabuchi 1986, Lauri 1988, Sveikauskas et al 1988). The estimates of urbanization economies range from 0.01 to 0.20, but the majority of values are under 0.10. This indicates that a doubling of city size is typically associated with an increase in productivity of somewhere between 1% and 10%. The estimates given in the table above are all positive although Henderson (1986, 2003a) does report difficulties in identifying urbanization effects on productivity. It is worth pointing out that of the studies shown in table 1 only those by Henderson (1986, 2003a) and Nakamura (1985) treat urbanization and localization economies within the same estimating equation. They are able to do this by estimating an industry level production function for some particular industrial sector in contrast to the other studies of urbanization where estimation is based on an aggregate function for all manufacturing. Table 1 shows four estimates oflocalization econoµiies. Nakamura (1985) estimates the effect oflocalization economies on the productivity of 20 manufacturing industries. He quotes an unweighted average elasticity of productivity with respect to industry size of0.05. This compares to an average city size elasticity of0.03 and thus Nakamura concludes that the effects oflocalization tend to be more significant than those of urbanization. Henderson (1986) also finds weak evidence of urbanization economies using industry level data for US MSAs and Brazilian cities but does find substantial evidence of localization economies. His estimates of localization economies for Brazil vary by industry, with a maximum elasticity estimate of0.20 and a minimum of 0.03, the mean value over 10 industries is 0.11. For US MSAs Henderson (1986) again finds substantial evidence of localization with a range in estimated elasticities of 0.09 to 0.45 for selected industries and a mean value of 0 .19. He concludes that economies of agglomeration tend to be ones of localization not urbanization and that localization economies tended to be strongest in the sectors in which cities specialise but that they diminish as city size increases. Henderson (2003a) finds similar results. He estimates plant level production functions for high-tech and machinery industries in the US using a variable recording the number of own industry plants to test for localization economies. He finds that localization effects are strong for high tech industries but not for machinery. . In addition to studies using MSA population and employment to construct variables representing city and industry size there are those that have incorporated some measures of distance or density into the specification of agglomeration effects. Two recent papers are particularly interesting in this respect. First, is the study of state level labour productivity and the density of economic activity by Ciccone and Hall (1996). They develop two spatial economic models; one based on the neo-classical conception that density can affect productivity through local geographical externalities, another which emphasises diversity of 1 The studies by Aaberg (1973), Kawashima (1975), and Moomaw (1981) proxy capital using measure of non-labour income, those by Fogarty and Garofalo (1978) and Sveikauskas et al (1988) create capital data through the perpetual inventory method, while those by Sveikauskas (1975), Moomaw (1985) and Lauri (1988) derive estimating equations which avoid the need for capital data altogether. Land and Urban Policies for Porvety Reduction 101 intermediate services where spatial density gives rise to aggregate increasing returns. From these models they derive an equation to estimate the effects of county-level employment density on aggregate state productivity. They find that over 50% of the variance in aggregate labour productivity across states can be explained by variance in the density of employment and that a doubling of employment density is associated with a 6% increase in average labour productivity. Ciccone (2002) extends the analysis to European data and estimates an elasticity of labour productivity with respect to employment density of 0.045. Second, is the paper by Rosenthal and Strange (2003) which used distance based measures at the establishment level to test for the extent of geographical externalities. Using the zipcode of the establishment as a centroid they construct distance rings at 1 mile, 5 miles, 10 miles and 15 miles. For the six industries they study that find that localization economies are present but that the strength of these decreases rapidly across space and substantially even within a five mile radius of the plant. Regarding urbanization economies they identify relatively small and inconsistent effects. Density and distance based measures have been used by other researchers. Fogarty and Garofalo (1988) estimate a production function with a vector of agglomeration effects that includes manufacturing employment density. They show that density has a strong positive non linear effect on productivity and that a change in spatial distribution of the density of industry may affect productivity substantially. Henderson et al (1995) estimate a growth model which finds externalities positively associated with own industry employment concentration. Hansen (1990) estimates a production function with distance based agglomeration measures to explore the trade off between factor costs and productivity in the Sao Paulo region of Brazil. He finds that productivity is enhanced by agglomeration as represented by distance to the centre of Sao Paulo but that there is a trade off because costs diminish with distance. Hanson (1996a, 1996b, 1997) explores relation- ships between agglomeration, productivity and wages for the garment sector of Mexico City. His data support the existence of localization economies and show that regional wages decrease by distance from the centre of Mexico City. Duranton and Overman (2002) develop distance based tests of localization for the UK. They find that 51%of4 digit industries are genuinely localised at an acceptable statistical level and that localization takes place at small scales, mostly below 50 kilometres. The role of transport investment In terms of agglomeration public infrastructure can be treated as an unpaid factor of production. Positive spatial externalities exist when an urban area provides an input that lowers costs for firms. If costs are lowered for only one industry we have localization economies. If costs are lowered for all firms we have urbanization economies. Eberts and McMillen (1999) cite two bodies of literature; one that is concerned with the impact of agglomeration .on productivity, and one that is concerned with the impact of infrastructure provision on productivity. Analytically these two topics have remained largely distinct, due in part to difficulties in constructing adequate data on public infrastructure at the metropolitan level. A number of studies have estimated the impact of infrastructure investment on productivity at the national or state levels. (see Gramlich 1994, Holtz-Eakin 1994, Fernald 1999, Rovolis and Spence 2002 for reviews of this literature). Some authors have found positive effects on productivity from public infrastructure provision (e.g. Aschauer 1987, 1988, 1989, 1990, Munnell 1990a, 1990b, Lynde and Richmond 1992, 1993a, 1993b, 104 Land and Urban Policies for Porvety Reduction In addition to the firm level commercial data we make use of official government data to represent agglomeration. The Annual Business Inquiry (ABI) provides information on the number of jobs by industry for a variety of geographical bases including wards. These data can be used to represent city size and industry scale. The ABI data are the most detailed industrial and geographical economic data available for Britain. These data give the number of employees in employment broken down by the 1992 SIC (defining up to 504 industrial sectors). The model. The translog production function We model the effects of agglomeration economies within the framework of a production function. Let the production function for the firm be Y =g (z)f(X) (1) where Yis the output level ~f the firm, Xis a vector of factor inputs with elements Xi (i = 1, ... ,n), and g(z) is a vector of influences on production which are Hicks' neutral in nature including those that arise from the firm's 'environment' such as agglomeration economies. If inputs are rented in competitive markets the first-order conditions for output maximisation subject to an expenditure constraint are (2) where Wi is the price of the ith input, and A, is a Lagrange multiplier which is the reciprocal of marginal cost "CI "Y. The expenditure constraint is given by, "W.X.=C, £..J I 1 (3) where C is total cost. From (2) and (3) /l.,= Li(aY1axi)xi (4) c and substituting (4) back into (2) after rearrangement yields the inverse input demand equations w. , ay;ax ( ) -t=L;(oYloX;)X;-g x (S) Note that these inverse input demand functions determine prices as functions of quantities as opposed to ordinary demand functions which determine quantities in terms of prices. Equation (5) can be written in cost share form (sf) as Land and Urban Policies for Porvety Reduction 105 Sc= W; X; = aIn YI aIn X; . c L aIn YI aIn xi (6) The production function described in equation (1) can be represented by the translog approximation. i 1 i i lnY=a 0 +g(z)+ L ailnXi+- LL yijlnX;lnXJ (7) i=l 2;= j= Given (6) appropriate differentiation of (7) yields the cost share equations. a;+ LY;;lnX i Sc= i ' La;+°L.LY j ; j lnX 1 (8) The translog parameters can be efficiently estimated by simultaneously estimating (7) and (8) as a nonlinear multivariate regression system. The specification of agglomeration economies From the literature review given above we know that previous research has typically used total metropolitan population or employment to provide an empirical measure of city size and total metropolitan employment in some industry i to represent the size of that industry. Such simple measures of agglomeration are not available for Britain. There are no good sources of data for British metropolitan areas and the aggregate data that do exist are for administrative areas that do not readily correspond to 'cities'. Perhaps more importantly it could be argued that in a small island country such as Britain it is hard to define distinct metropolitan areas. For instance, while Greater Manchester and Liverpool are nominally two separate cities, there is interaction between the two over relatively small distances that arguably prevents them from being truly distinct. Likewise it is conceivable that a firm located outside the London conurbation can still enjoy agglomeration benefits through proximity that arise from the scale of London and its industries. The point is that in a small country like Britain we can legitimately ask where the actual influences from urban centres ends. For this reason we model agglomeration economies using a measure that incorporate both proximity and the scale of economic activity and that can be calculated for very small areas throughout the country. Specifically, we use ward level employment data to construct a measure of accessibility experienced by each firm in the FAME data. The density of economic activity, or agglomeration, experienced by any firm in industry o located in ward is given by -a) U;,=[vf(A,;n)j" + f E. i*-J ( Ei6dij' (9) where Ei is total employment in ward i, Ai is the area of ward i, Ei is total employment in ward j, and dii is the 0 distance between i and j. The value ofo determines the effect of distance on the strength of externalities for each industry o. 106 Land and Urban Policies for Porvety Reduction Results The translog estimations are conducted for 20 industry groups consisting of 8 service activities and 12 manufacturing activities. The service activities investigated are. 1. Finance & insurance (SICs 65, 66 and 67) 2. Real estate activities (SIC 70) 3. Computer and related activities (SIC 72) 4. Business and management consultancy activities (SIC 7414) 5. Architecture and engineering activities (SIC 742) 6. Adverti'sing (SIC 744) 7. Labour recruitment and provision of personnel (SIC 745) 8. Motion picture and video activities, radio and television (SICs 921 and 922). The manufacturing industries are. 1. Manufacture of food products and beverages (SIC 15) 2. Manufacture of textiles, wearing apparel, dying and dressing of fur (SI Cs 17 and 18) 3. Manufacture of wood and wood products (SIC 20) 4. Manufacture of pulp, paper and paper products (SIC 21) 5. Publishing, printing and reproduction of recorded media (SIC 22) 6. Manufacture of chemical and chemical products (SIC 24) 7. Manufacture of rubber and plastic products (SIC 25) 8. Manufacture of basic metals and fabricated metal products (SI Cs 27 and 28) 9. Manufacture of office machinery and computers (SIC 30) 10. Manufacture of radio, television and communication equipment (SIC 32) 11. Manufacture of medical, precision and optical instruments, watches and clocks (SIC 33) 12. Manufacture of motor vehicles and transport equipment (SI Cs 34 and 35) The choice of manufacturing industries has been made largely on the basis of data availability; we have excluded industries for which we have an insufficient number of firms for estimation. For services, we have excluded sectors such as retail, education, health, and public administration which are less interesting from the point of view of agglomeration. Results for service industries are shown in table 1 below. For brevity we here present only the elasticity results relating to the agglomeration term which we denote Pa.Full translog results can be found in Graham (2005). Land and Urban Policies for Porvety Reduction 107 Table 1: Estimates of the elasticity of output with respect to agglomeration externalities for service sectors from trans log functions. (1) (2) (3) (4) (5) (6) (7) (8) R. 0.116** 0.130** 0.072** 0.176** 0.061 ** 0.101 ** 0.000* 0.256** Pa (0.015) (0.025) (0.023) (0.033) (0.021) (0.011) (0.000) (0.067) Standard errors in parentheses The estimates shown in table 1 indicate strongly that there is a positive association between agglomeration and higher productivity. Estimates range from 0.072 to 0.256. The strongest effects are found in motion picture video & TV activities (8), business and management consultancy services (4), finance and insurance (1), and real estate activities (2). Weaker, but still positive and significant effects are found in advertising (6), computer activities (3), architecture and engineering (5)and labour recruitment (7). For the manufacturing industries we found that the translog regressions performed poorly. For many sectors samples are relatively small yielding inadequate variance to implement the full translog model, including quadratic and interactive terms. For this reason, we decide to pool the industry data and estimate a Cobb- Douglas equation using dummy variables to distinguish our separate industry groups. We hypothesise that these dummy variables will capture a variety of industry specific differences that give rise to variation in productivity including differences in RTS. We use the Cobb-Douglas form here because for some se~tors we have insufficient cost data to estimate the full translog system. The estimating equation is y =ex+ g (-) + /3 K ln (]() ln L L + ( /3 K + /3 L - 1) ln L + £ . (10) where in addition to variables already defined D is a matrix of dummy variables that differentiates by industry type, K is capital input and L is labour input. 108 Land and Urban Policies for Porvety Reduction Results are presented in table 2 below. Table 2: Cobb-Douglas estimates for manufacturing firms with a single accessibility measure of agglomeration. S/C/5 3.960** (0.360) S!C/718 3.636** (0.367) SIC20 3.825** (0.363) SIC21 3.552** (0.365) SIC22 3.549** (0.370) SIC24 3.719** (0.365) SIC25 3.511 ** (0.365) SIC2728 3.511** (0.361) SICJO 3.627** (0.367) SIC32 3.662** (0.367) SIC33 3.409** (0.364) SIC3435 3.734** (0.363) pk 0.608** (0.011) (f3k .+f31 -1) -0.005 (0.011) [3£ 0.094** (0.027) R1 0.581 n 2331 Standard errors in parentheses The dummy variables for industry group are all significant at 1% confirming the need to differentiate the data in this way. The estimate of Wk f PL -1) is statistically insignificant indicating that we cannot reject the hypothesis that there are CRS. Thus, the elasticity of output with respect to capital is 0.608 and the elasticity of output with respect to labour is 0.392. The coefficient associated with the accessibility agglomeration variable is positive and significant at the 1% level. Across our sample, holding factor inputs and industry specific effects constant, we find that a 10% increase in the effective density of jobs available to the firm is associated with a 0.94% increase in productivity. Thus, even for the manufacturing sectors, using a single accessibility measure, we find evidence of agglomeration or density externalities. We have seen from estimates based on our single accessibility measures that there appear to be positive agglomeration, or density, externalities for service sector and manufacturing firms. One important issue is that we have not able to control for the particular function of firms or for the skill characteristics of the labour inputs of firms. We have classified our firms according to the SIC, but yet we know that data classified by industry may not contain homogenous activities. Firms differ in the functions they perform, even within the same detailed SIC, and there is evidence to show that this 'functional specialisation' is related to city size For instance, Duranton and Puga (forthcoming), show that while the largest cities tend to specialise in functions related to business services and management, smaller cities tend to specialise in production based activities. Rice and Venables (2004), analysing earnings in relation to productivity at the NUTS3 level, find that Land and Urban Policies for Porvety Reduction 109 occupational structure is positively correlated with productivity and therefore that areas with higher productivity also tend to have employment structures in high paying occupations 2 • Therefore, it is legitimate to ask whether we have identified real productivity differences between homogenous firms due to agglomeration externalities, or whether instead we have found a kind of 'functional gradient' with firms in the most urbanised locations performing different and more productive types of activities. This is, however, a very difficult question to address empirically. We do not have information on the functions our firms perform or on the occupational structure of their workforces. What we do know is that if there is a 'functional gradient' it is likely to be at a maximum in London. It may therefore be informative to extract London firms from our sample and see how estimates change with London excluded. Tables 3 and 4 show results for service and manufacturing firms, excluding those with a London location based on the single accessibility measure of agglomeration (equation 9). Table 3: Estimates of the elasticity of output with respect to agglomeration externalities for service sector firms outside London from translog functions with a single accessibility measure of agglomeration. (1) (2) (3) (4) (5) (6) (7) (8) A 0.284** 0.136* 0.094** 0.252** 0.094** 0.274** 0.047 0.261 1-'a (0.469) (0.059) (0.036) (0.062) (0.040) (0.070) (0.121) (0.181) Standard errors in parentheses For six of our eight service sectors we are still able to identify a positive density externality in the data having excluded London from the sample. Furthermore, for each of these industries we find that we actually obtain higher estimates using the sample that excludes London. Table 4 presents the estimates for manufacturing firms based outside London 2 Rice and Venables (2004) go on to analyse relationships between proximity to economic mass and the productivity and occupational components of variation in average earnings. Interestingly they find a robust relationship between productivity and proximity to economic mass but not between occupational composition and proximity to economic mass. 110 Land and Urban Policies for Porvety Reduction Table 4: Cobb-Douglas estimates for manufacturing firms outside London with a single accessibility measure of agglomeration. SIC15 3.870** (0.560) SIC1718 3.426 * * (0.570) SIC20 3.690 ** (0.563) SIC21 3.422 ** (0.569) SIC22 3.448 ** (0.566) SIC24 3.605 ** (0.568) SIC25 3.458 ** (0.569) 51(2728 3.385 ** (0.568) SIC30 3.478 ** (0.569) SIC32 3.544 ** (0.570) SIC33 3.405 ** (0.568) SIC3435 3.599 ** (0.568) pk 0.608 ** (0.011) (pk .+?Pl -1) -0.017 (0.010) P, 0.099** (0.041) Ri 0.632 n 1986 Standard errors in parentheses The estimate of the elasticity of output with respect our accessibility index is almost identical for manufacturing firms using either the sample including or excluding London, approximately 0.1. Thus, for manufacturing firms outside London we still find evidence of positive agglomeration or density externalities. Of most interest from the results given above is that for some service sectors we find the magnitude of the agglomeration estimates to be consistently higher when London is excluded. For Advertising (744) and finance and insurance (SI Cs 65 to 67) the estimates are over two times as large with London firms excluded. Similarly, real estate activities (SIC 70), computer services (SIC 72), architecture and engineering (SIC 742), and business and management consultancy services (SIC 7414) all have higher estimates based on the sample that excludes London firms. Why might we find these increases in the magnitude of estimates when using samples that excludes London firms? One possible explanation could relate to the sort of decomposition that Venables and Rice (2004) make regarding productivity and occupation. If it is the case that the occupational (rather than productivity) contribution to wages is particularly high in London, and that elsewhere in the country productivity plays a larger role, then the exclusion of London could lead to a more pronounced productivity-proximity relationship. This is because the exclusion of London,firms reduces the 'explanation' for productivity differences that are based on occupational mix. Land and Urban Policies for Porvety Reduction 111 A second more straightforward possibility for the increased magnitude of effective density estimates in the sample excluding London is the existence of some diseconomies of city size in the capital. We know, for instance, that congestion is much worse in London than in other British towns and cities and this could impact upon the efficiency of firms. In particular, the effect of congestion is such that it increases the generalised cost of travelling some distance. This means that our measure of effective density, which is based on distance and not generalised cost, in reality, contains measurement error because it does not account for congestion. Distance travelled in London costs more than in other cities or towns of the UK. Thus, it is possibie that we are overestimating the effective density of high density locations by not accounting for the fact that their transport networks are often more congested. As a result the elasticities may be biased downwards when London is included. . The third possibility is simply that diminishing returns to density set in at some stage such that on the margin the effect of productivity by increasing density is lower in London than elsewhere. It is hard to reach any firm conclusions comparing estimates obtained using the full sample of firms to those based on data that excludes London firms. The evidence is mixed and in many cases the substantial drop in the number ofobservations poses problems in drawing any meaningful comparisons between the two samples. What we can say, however, is that for many of our sectors positive density externalities can still be identified when London is excluded and that there does therefore appear to be a strong productivity-density gradient for the rest of the country. Conclusions In this paper we have estimated the effects of 'city size' on productivity for different sectors of the UK economy. Our specification of 'city size' is based on a measure of accessibility which recognises the importance of both the scale and proximity of economic activity. It is a measure that contains an implicit transport dimension by capturing the effective density that is available to firms. Our results provide compelling evidence of a strong association between our measure of ~ity size and productivity for both service and manufacturing sectors of the UK economy. Thus, if transport investment increases the effective density of locations there could be an effect on productivity which can be quantified for economic appraisal in the way suggested by Venables (2003). On the other hand, our results also indicate that there may be a gradient of functional spe:cialisation at work in creating productivity differentials and that this gradient may be correlated with city size. If so, we have to caution a simplistic view which would imagine productivity differences amongst homogenous activities distributed across the urban hierarchy with those in the largest cities being the most efficient. Unfortunately, the limitations of the SIC do not allow us to analyse this issue in great depth. What we can say, is that even if the productivity differences are based on functional specialisation, we may still expect external benefits from transport improvements if they assist the process of functional. For nations in the process of rapid urbanization one implication is that urban growth may offer opportunities to capture the benefits of externalities in the form of productivity gains. Transport investments can help to foster the conditions for city size to make an impact on the efficiency of firms. Understanding the relationships between city size, transport improvements and productivity can also be helpful in making the economic case for transport investment. 112 Land and Urban Policies for Porvety Reduction Annex: Unit cost, Commuting cost benefit w+-~~---,-~~~~,;<--~~~-w Wage gap Rent 0 l C-cost A E Number of woekers Figure 1: Urban equilibrium Unit cost, New Commuting benefit cost 0 E Number of workers Figure 2: Net gains from transport improvement New Commuting cost Unit cost, benefit 0 Number of workers Figure 3: Net gains from transport improvement with endogenous productivity Land and Urban Policies for Porvety Reduction 113 References Aaberg Y 1973. 'Regional productivity differences in Swedish manufacturing', Regional and Urban Economics, 3, 131-156. Aschauer DA 1987. 'Is government spending simulative?', Federal Reserve Bank of Chicago, Staff Memorandum. Aschauer DA 1988. 'Government spending and the falling rate of profit', Economic Perspectives, May I June, 17-25. Aschauer DA 1989. 'Is public expenditure productive?', Journal of Monetary Economics, 23, 177-200. Aschauer DA 1990. 'Why is infrastructure important?', in Munnell A (ed) Is there a shortfall in public investment? Federal Reserve Bank of Boston. Ciccone A and Hall RE 1996 .. 'Productivity and the density of economic activity', American Economic Review, 86, 54-70. Daltenberg D 1987. 'Estimates of elasticities of substitution between public and private inputs in the manufacturing sector of metropolitan areas' PhD dissertation (University of Oregon). Deno KT 1988. 'The effect of public capital on U.S. manufacturing activity', Southern Economic Journal, 55, 400-411. Duranton G and Overman H 2002. 'Testing for localization using micro-geographic data', CEPR discussion paper 3379. Duran ton G and Puga D forthcoming. 'From sectoral to functional urban specialisation' Journal of Urban Economics Eberts RW 1986. 'Estimating the contribution of urban public infrastructure to regional economic growth', Working Paper no. 8620(Federal Reserve Bank of Cleveland). Eberts RW and McMillen DP 1999. '.Agglomeration economies and urban public infra-structure' in P Cheshire and ES Mills (eds) Handbook of regional and urban economics, Volume III, 1455-1495. North Holland: New York. Fernald JG 1999. 'Roads to prosperity? Assessing the link between public capital and productivity', American Economic Review, 89, 619-638. Fogarty MS and Garofalo GA 1978. 'Environmental quality income trade-off functions with policy applications', paper presented at the Southern Regional Science Association Meeting. Fogarty MS and Garofalo GA 1988. 'Urban spatial structure and productivity growth in the manufacturing sector of cities', Journal of Urban Economics, 23, 60-70. Gerking S 1994. 'Measuring productivity growth in US regions: a survey', International Regional Science Review, 16, 155-186. Graham DJ 2005. Wider economic benefits for transport improvements: link between city size and productivity: Stage 1 report, Department for Transport, London. 114 Land and Urban Policies for Porvety Reduction Graham DJ 2006. Wider economic benefits for transport improvements: link between city size and productivity: Stage 1 report, Department for Transport, London. Graham DJ forthcoming. 'Agglomeration, productivity and transport investment', Journal of Transport Economics and Policy. Gramlich EM 1994. 'Infrastructure investment: a review essay', Journal of Economic Lit~rature, 32, 1176-1196. Hansen ER 1990. 'Agglomeration economies and industrial decentralisation: the wage-productivity trade- off', Journal of Urban Economics, 28, 140-159. Hanson GH 1996a. 'Localization economies, vertical organisation and trade', American Economic Review, 86, 1266-1278. Hanson GH 1996b. 'Agglomeration, dispersion and the pioneer firm', Journal of Urban Economics, 39, 255-281. Hanson GH 1997. 'Increasing returns, trade and the regional structure of wages', The Economic Journal, 107, 113-133. Haughwout AF 1999. 'State infrastructure and the geography of employment', Growth and Change, 30, 549-566. Henderson JV 1986. 'Efficiency of resource usage and city size', Journal of Urban Economics, 19, 47-70. Henderson JV 1988. Urban development: theory, fact and illusion. Oxford University Press: Oxford. Henderson JV 2003a. 'Marshall's scale economies', Journal of Urban Economics, 53, 1-28. Henderson JV 2003b. 'The urbanization process and economic growth: the so-what question', Journal of Economic Growth, 8, 47-71. Henderson JV, Kuncoro A and Turner M 1995. 'Industrial development in cities', Journal of Political Economy, 103, 1067-1085. Holl A 2004a. 'Manufacturing location and impacts of road transport infrastructure: empirical evidence from Spain', Regional Science and Urban Economics, 34, 341-363. Holl A 2004b. 'Transport infrastructure, agglomeration economies, and firm birth. Empirical evidence from Portugal, Journal of Regional Science (forthcoming). Holtz-Eakin D l 993a. 'New Federal spending for infrastructure: should we let this genie out of the bottle', in Public Infrastructure Investment: A Bridge to Productivity Growth?, The Jerome Levy Economics Institute, Bard College, NY Holtz-Eakin D 1993b. 'State specific estimates of state and local government capital', Regional Science and Urban Economics, 23, 185-210. Holtz-Eakin D 1994. 'Public-sector capital and the productivity puzzle', Review of Economics and Statistics, 76, 12-21. Kawashima T 1975. 'Urban agglomeration economies in manufacturing industries', Papers of the Regional Science Association, 34, 157-175. Land and Urban Policies for Porvety Reduction 115 Kim HY 1992. 'The translog production function and variable returns to scale', Review of Economics and Statistics, 74, 546-552. Moomaw RL 1981. 'Productivity and city size: a review of the evidence', Quarterly Journal of Economics, 96, 675-688. Moomaw RL 1983a. 'Spatial productivity variations in manufacturing: a critical survey of cross sectional analyses', International Regional Science Review, 8, 1 - 22. Moomaw RL 1983b. 'Is population scale a worthless surrogate for business agglomeration economies', Regi- onal Science and Urban Economics, 13, 525-545. Moomaw RL 1985. 'Firm location and city size: reduced productivity advantages as a factor in the decline of manufacturing in urban areas', Journal of Urban Economics, 17, 73-89. Munnell AH 1990a. 'Why has productivity growth dec;lined?', Productivity and Public Investment, New England Economic Review, January I February, 3-22. Munnell AH 1990b. 'How does public infrastructure affect regional economic performance?' in Munnell A (ed) Is there a shortfall in public investment? Federal Reserve Bank of Boston. Nakamura R 1985.'Agglomeration economies in urban manufacturing industries: a case of Japanese cities', Journal of Urban Economics, 17, 108-124. Rice P and Venables AJ 2004. 'Spatial determinants of productivity: analysis for the regions of Great Britain', Centre for Economic Performance Working Paper Series, Paper No' CEPDP0642. Rosenthal Sand Strange W 2003. 'Geography, industrial organisation and agglomeration', Review of Economics and Statistics, 85, 377-393. Rosenthal Sand Strange W 2004. 'Evidence on the nature and source of agglomeration economies' forthcoming in V Henderson and J Thisse (eds) Handbook of Urban and Regional Economics, Vol 4. (http:/ I econ. pstc. brown.ed u/ facul ty/henderson/WillAndStuart. pdf) Seitz H 1993a. 1\. dual economic analysis of the benefits of the road network', Annals of Regional Science, 27, 223-239. Seitz H 1993 b 'The impact of the provision of urban infrastructures on the manufacturing industry in cities', paper presented at the 33rd European Congress of the Regional Science Association, Moscow, Russia. Seitz H 1994. 'Public capital and the demand for private inputs', Journal of Public Economics, 54, 287-307. Seitz H 1995. 'The productivity and supply of urban infrastructures', Annals of Regional Science, 29, 121-141. Shefer D 1973. 'Localization economies in SMSAs: a production function analysis', Journal bf Regional Science, 13, 55-64. Sveikauskas LA 1975. 'The productivity of cities', Quarterly Journal of Economics, 89, 393-414. Sveikauskas L, Gowdy J and Funk M 1988. 'Urban productivity: city size or industry size', Journal of Regio- nal Science, 28, 185-202. Venables A 2003. Productivity effects of transport improvements, working paper, Lon-don School of Economics. IMPACT OF TRANSPORT INFRASTRUCTURE & SERVICES ON URBAN POVERTY AND LAND DEVELOPMENT: A CASE STUDY - COLOMBO, SRI LANKA Amal S. Kumarage* Abstract Colombo is a relatively small city with a resident population of 700,000 with a day time inflow of a million persons. Its area is 3,730 hectares. The Colombo Metropolitan Region (CMR) which serves as the suburban feeder area for Colombo city has a population of over 4.6 million with a gross population density of 13 persons per hectare. In the City of Colombo the density is 188 persons per hectare. The land use distribution in City of Colombo shows that residential use takes up 40%, of the available land, while transport & communications takes up 13%, with a further 30% presently developed for commercial and administrative purposes, with around 17% land bare or still under non-urban use. The residential densities within the city range from between 165 to 1,537 persons per hectare. The highest densities are accompanied by concentrations of people living in illegal squatter settlements that are badly over crowded with respect to facilities available within them. These have however become popular forms of settlements for the poor in the absence of affordable public or private sector housing programs. It is estimated that at present about 35% of the city's population lives in these settlements, which have semi permanent houses, shared toilets and poor sanitation conditions. This shortage of housing for the poorest sections of the city is commonly attributed to economic indictors particularly affordability to the low income consumer to purchase or rent, scarcity ofland and high land prices and high construction costs. The paper examines a factor that is usually overlooked -, i.e. transport -and how inappropriate transport facilities contribute to the demand for these central-city informal settlements rather than those of suburban areas where better services are available. The paper is based on two sources of data and their respective analysis. In the first instance, the paper examines briefly the development of the form of the city and the growth in commuter traffic. It traces the legacy of urbanization dating back to the 16th century centered on the development of the Port of Colombo under Portuguese occupation. Under British occupation in 1871, the City had an extent of 2,449 hectares with a population of 98,847 persons. The density doubled by 1931 by which time the city had grown to 3,368 hectares with a population of284,155, largely due to annexation of surrounding areas. This density doubled by 1981, by which time the land area had reached a near maximum of 3,711 hectares. The most recent land-use strategic plan has proposed to reduce the extent of residential land use from 1,401 hectares to 691 hectares by 2010 to provide for more commercial development At the same time, traffic flows crossing the city boundary increased at the rate of2.8% per annum during the period 1961 to 1979. However it has increased at a much higher rate of 5.4% per annum over the last two decades. The passenger growth observed during the period 1985-95 was 4.7%, with bus transport growth at * Professor, University of Moratuwa, Sri Lanka I Chairman, National Transport Commission, Sri Lanka I kumarage@sltnet.lk 118 Land and Urban Policies for Porvety Reduction 4%, private vehicles growing at 11.8% and railways at 2.8%. It analyses the fact that these growth rates are inversely proportional to the cost of travel. In other words, the cheapest forms have had the lowest growth. In all, there are presently an estimated 2 million passenger crossings (both directions) per day in 315 ,504 vehicles of which 80% are private vehicles. The second source of data is from a survey of the reasons why the poor continue to live in these illegal squatter settlements. The ability to access work is given as a major reason for putting up with such an environment. The paper also compares the quality oflife of people who are engaged in the same occupation but live within and outside the city. The comparison is based on the relative costs of transport, time of travel, availability of late night travel and social parameters such as type of housing, status of children's education, etc. The survey also investigates the impact of transport provision as a means of livelihood of the urban poor. This is particularly significant with respect to three-wheeler (auto-taxi) drivers who are residents in these settlements and prefer to live close to the city centre which is a focal point of their work and cannot drive their vehicles long distances for the night. It is estimated that around 25,000 to 30,000 new houses would be required to house these low income families. The land that is presently occupied by these settlements can be used partially for this purpose. However, most resettlements would have to take place outside the city. The land values in Colombo City during the period 1985 to 1998 have increased at the rate of 16.5% per annum (p.a.) in nominal terms and adjusted for inflation this is approximately 5% p.a. while that of the suburban areas in the ranges of between 10 to 20 kms from the centre has also increased by around 18% p.a. so that the real increase adjusted for inflation is around 6.5% p.a.This makes purchase ofland nearly impossible for poor people. The alternative areas for relocation identified as Kesbewa, Maharagama, Kaduwela, Gampaha and Panadaura, which are all suburban centers, are located at distances of between 20 to 30 kms from the city centre. The relocation of the poor to these locations will make accessing jobs in the city more difficult for them. Ir is most unlikely that they will move since it adversely affects their livelihood. The survey by tracing the location of work, alternative housing locations, access and cost of transport etc finds that land prices in suburban areas which are alternative locations for the urban poor to be relocated are usually away from the main transport corridors and are presently poorly served by public transport. The irregular hours that the poor work are not conducive to public transport which usually operates well only during peak periods. Thus, the costs of reaching these alternative areas are high and hence, the need to reside in the city. This increases the value ofland and also overcrowding in settlement areas which are the only such affordable lands for the poor. In addition this puts pressure on services in urban areas and results in the poor not having adequate equal access to these services, which are more freely available in suburban areas. For example, the city has the most popular schools, bur the ones attended by the children of the poor are neglected when compared to similar schools in suburban areas. Similarly, incidences of health and safety are higher; the involvement in crime and other related activities is also higher in these squatter settlements. The research concludes that developing transport infrastructure and services to enable the poor to access employment available within the city is an important aspect of dealing with the quality of life of the urban poor. It also establishes a link between transport development and land development especially with respect to prices and affordability of urban housing. In order to ensure that the quality oflife of the poor is improved, Land and Urban Policies for Porvety Reduction 119 the paper proceeds to give a checklist of questions that should be considered in the development of a city which will address the transport related problems faced by the low income households. Furthermore the paper concludes with options for integrating transport as a means of land development in urban centers taking Colombo as a case study. Introduction The City of Colombo serves as the 'Primate City' in modern Sri Lanka. Colombo and its metropolitan area-referred to as the Colombo Metropolitan Region (CMR)-fall within the Western Province, which is the most densely populated and economically active region within the country Table 1. Transportation activity within this region is also the densest in Sri Lanka. Table 1: Summary of Vital Statistics of Colombo Metropolitan Region CMR Sri Lanka Percentage (%) Land Area (sq. km.) 3,593 62, 705 5.8 Population (2001- Mn) 5,361 18, 732 28.6 GDP (1994- Rs. Mn) 1 22, 582 51,227 44.1 Vehicle Licenses (2001) 456, 164 955,238 47.7 Sea Freight (2001) TEU 1, 726,605 N/A Air Traffic (Pax. Movements-2001) 2,916.407 2,916,407 100.0 History: From ancient times, Sri Lanka has been largely an agricultural economy. In recent history, particularly under colonial rule, the development of the Port of Colombo and the availability of suitable human resources led to the majority of industries locating within one hour travel distance from the port. The growth of industries and the development of Colombo as the administrative capital and primary commercial center of the country have formed the basis of the physical expansion of Colombo and its environs. . The legacy of urbanization dating back to the 16th century centered on the development of the Port of Colombo under Portuguese occupation. Under British occupation in 1871, the City had an extent of 2,449 hectares with a population of 98,847 persons. The density doubled by 1931 by which time the city grew to 3,368 hectares with population growing to 284, 155 largely due to annexation of surrounding areas. This density doubled by 1981, by which time the land area had reached a near maximum of 3,711 hectares. The most recent land use strategic plan has proposed to reduce the extent of residential land use from 1,401 hectares to 691 hectares by 2010 in order to provide for more commercial development (UDA, 1998). Geographic: Colombo is a relatively small city with a resident population of around 700,000 with a day time inflow of a million persons. Its area is 3,730 hectares. The Colombo Metropolitan Region (CMR) which serves as the suburban feeder area for Colombo city has a population of over 5 .3 million with a gross population density of 15 persons per hectare. In the City of Colombo itself the density is 188 persons per hectare. 1 In 1994, 1 US $was Rs 60, while it is Rs 100 in 2005. 120 Land and Urban Policies for Porvety Reduction Table 2: Population (2001) Area Population 2001 Colombo Municipal Area 697,396 Colombo District 2,234,289 Colombo Metropolitan Region 5,361, 185 Sri Lanka 18, 732,255 Demographic: The land use distribution in City of Colombo shows that residential use takes up 40%, of the available land, while transport & communications takes up 13%, with a further 30% presently developed for commercial and administrative purposes, with around 17% land bare or still under non-urban use. The residential densities within the city range from between 165 to 1,537 persons per hectare (UDA, 1998). The highest densities are accompanied by concentrations of people living in illegal squatter settlements that are badly over crowded with respect to facilities available within them. These have, however, become popular forms of settlements for the poor in the absence of affordable public or private sector housing programs. It is estimated that at present about 35% of the city's population lives in these settlements, which have semi permanent houses, shared toilets and poor sanitation conditions. This shortage of housing for the poorest sections of the city is commonly attributed to economic indictors particularly Figure 1: Sri Lanka affordability to the low income consumer to purchase or rent, scarcity of land and high land prices and high construction costs. Transport: During the period 1961to1979, the traffic flows crossing the city boundary increased at the rate of 2.8% per annum. However it has increas~d at a much higher rate of 5.4% per annum over the last two decades. The passenger growth observed during the period 1985-95 was 4.7%, with bus transport growth at 4%, private vehicles growing at 11.8% and railways at 2.8%. It analyses the fact that these growth rates are inversely proportional to the cost of travel. In other words, the cheapest forms have had the lowest growth. In all, there are presently an estimated 2 million passenger crossings (both directions) per day in 315,504 vehicles of which 80% are private vehicles (Kumarage, 2000). The desire lines which indicate the direction, distance and volume of flow arriving at the centre, for the commuting trips to Colombo City can be illustrated as in Figure 2. This shows that commuting trips are rather short distances, with a few exceptions, where low cost railway travel is available. Land and Urban Policies for Porvety Reduction 121 Housing: It is estimated that around 25,000 to 30,000 new houses would be required to house these low income families adequately. The land that is presently occupied by these settlements can be used partially for this purpose. However, most resettlements would have to take place outside the city. The land values in Colombo City during the period 1985 to 1998 have increased at the rate of 16.5% per annum (p.a.) in nominal terms and adjusted for inflation this is approximately 5% p.a. (UDA, 1998) while that of the suburban areas has also increased by around Figure 2: Commuting Desire Lines by Public Trasnport tot City of Colombo. 18%p.awheretherealrarewasaround6.5% ; _ p.a. This makes purchase of land nearly ! impossible for poor people. The alternative areas for relocation are located at distances between 20 to 30 kms from the city centre. The relocation of the poor to these locations will make accessing jobs in the city more difficult for them. It is most unlikely that they will move since it adversely affects their livelihood. Income: Income Distribution for the Western Province, as calculated from the Sri Lanka Integrated Study (1999/2000) data, is given in Table 3. This reinforces the position that two-thirds of the population is not engaged in income receiving occupations. It seems that Desire Line a significant proportion of income receiving count --1 (34%) fall within the lower half of income -2 range of up to Rs 3,0001= per month (US$ -3 -4 430), while 11 % falls in the income range of -5 :-s over Rs. 10,000/= (US$ 1,430) per month. I -1 Road Network Table 3: Income Distribution (1999/2000) 2 Income Range Western Province Sri Lanka Not employed/student/sick 66.1 64.9 Up to Rs 1,000/= 1.0 4.3 Rs 1,001 to Rs 2,000/= 4.2 6.2 Rs 2,001 to Rs 3,000/= 6.4 7.2 Rs 3,001 to Rs 5,000/= 9.8 8.6 Rs 5,001 to Rs 10,000/= 8.9 ·. 6.2 Rs 10,001 to Rs 25,000/= 2.4 1.8 More than Rs. 25,000/= 1.2 0.7 Total 100.0 100.0 2 In 1999/2000, the average conversion rate was 1 US $ = 70 Rs. 122 Land and Urban Policies for Porvety Reduction Objective & Scope of Paper The Sri Lanka Transport Sector Strategy Study (World Bank, 1997) notes that poverty alleviation requires a transport policy that is focused on the poor. The lack of such a policy and of relevant information has made it difficult to analyze how the transport sector is serving and helping the poor. It has been assumed that the mobility needs of the poor could be resolved by improving transport networks and public transport services in both rural and urban areas. Policies should address, among other things, the best ways to provide adequate and affordable access for the poor to get to work, particularly in rural and marginal urban areas, opportunities for generating employment through the transport sector, and the strategic use of transport to reduce regional disparities. There are no studies where the transport needs of the poor have been studied specifically. This paper examines the relationship between employment of the low income earners, their places of residence, and the transport linkages that are made available. Analysis of Income and Transport in Western Province This analysis is undertaken from aggregate socioeconomic data collected through Census and other household surveys and published from time to time. This data is not available for the City of Colombo. It does however exist for the Western Province. The objective of this analysis is to identify the patterns of (a) expenditure on transport and (b) of income of those living in the Western Province. Individual Income and Distance of Travel to Work Data from the Sri Lanka Integrated Survey (1999/2000) have been used to analyze the relationship between place of work and place of residence. Table 4 shows results for the Western Province (WP) compared to the rest of the country where over half of people working, do so within their own community. This could be interpreted in several ways. First, it might suggest that population is so distributed that the majority of the employment opportunities are located outside the communities they live in. Second, it might suggest a higher mobility for finding employment outside the local community, due to existence of acceptable transport services. Table 4: Relationship between Place of Work and Place of Residence Western Province Sri Lanka Same Community 51.2 66.0 Other Urban Community 37.3 23.9 Other Rural Community 0.6 0.8 Other 10.9 9.3 Total 100.0 100.0 Table 5 gives the cross-relationship between income and place of work/place of residence for the Western Province. These two tables show that there is a direct coaelation between individual incomes and the propensity to seek employment in other communities. This is an interesting phenomenon that could be due to the fact: Land and Urban Policies for Porvety Reduction 123 (a) That those who are able to commute outside their communities can get better incomes. (b) That those who have higher incomes tend to seek employment away from their own communities. Table 5: Individual Income and Place of Work with Respect to Place of Residence -WP Same Other Urban Other Rural Other 19 Community Community Community Not employed/student/sick 71.4 7.1 0 21.4 Rs 0 to Rs 1,000/= 76.2 9.5 4.8 9.5 Rs 1,001 to Rs 2,000/= 56.6 31.3 0 12.0 Rs 2,001 to Rs. 3,000/= 51.2 40.0 0 8.8 Rs 3,001 to Rs 5,000/= 44.9 45.9 0.5 8.7 Rs 5,001 to Rs 10,000/= 38.9 53.3 1.7 6.1 Rs 10,001 to Rs 25,000/= 54.2 35.4 0 10.4 More than Rs. 25,000/= 48.0 36.0 0 16.0 Total 50.6 38.3 0.6 10.4 In the case of (a) it relates to the availability and affordability of transport. This implies that poor transport will make people immobile and captive to their own communities, thus preventing them from accessing and holding employment that is higher paying. Both Tables 4 and 5 indicate that only those with incomes less than 1000/= per month appear to show a marked difference to other income categories with respect to the percentage of persons working within the same community. The amount of income that falls within this category in all probability refers to part time employment which cannot be compared with the full time employment as the commuting distances would be very much less in the case of the former. In the case of (b) above, it is a known social factor that higher paid employment is generally concentrated in centers (usually urban) and thus the average commuting distances would increase as people seek higher paying employment. This argument also can be used to explain why the percentage working in other urban areas increases with income and then begins to decrease when monthly incomes increase beyond Rs. 10,000/ =. This could possibly mean that relocation becomes more affordable when incomes are in that magnitude. The reverse inference of this observation is that when incomes are less than Rs 10,000/= per month, people are more likely to be constrained by the availability of transport facilities in seeking employment away from their community of residence. A comparison of the two tables indicates that in the Western Province, there is higher mobility between residence and employment communities for the same income groups. This means that people have to commute further as residential and employment areas tend to be more separated in urban and suburban areas. Occupation and Travel to Work Table 7 gives the cross-relationship between type of occupation and place of work/place of residence for the Western Province. There is relatively little mobility among those engaged in agriculture, as many people in this category are farming their own land or fishing, both activities generally being located close to residences. 124 Land and Urban Policies for Porvety Reduction Those in business, trade, and manufacturing activities also appear to be, in general, residing dose to their places of employment - for example, family-based businesses where home and shop or home and trade are located within the same premises. On the other hand, casual labour shows a somewhat higher propensity to seek employment in urban centers. These might be persons who are engaged in construction or similar work and who might not actually be commuting on a daily basis - more because of distance than transport fare. Salaried employees mostly travel outside their communities to urban communities for employment and show the highest degree of mobility. Table7: Type of Occupation and Place of Work with Respect to Place of Residence -WP Same Other Urban Other Rural Other Community Community Community Casual Labour 55.1 23.2 1.7 19.8 Salaried Employees 29.3 63.4 0.3 7.0 Busi ness!Trad e/M an ufa ctu ring 76.1 15.0 0.0 8.8 Personal Services 50.0 6.3 0.0 43.8 Agricultural 92.8 6.3 0.0 0.9 Income and Ownership of Vehicles Ownership of all types of vehicles in Western Province increases with income, as shown in Table 8. All income groups own bicycles in significant numbers and bicycles are the most common vehicle owned. Motorcycles are also used by all income groups, although their ownership levels become significant only when household incomes rise above Rs 5,000 per month. In the case of cars and vans, ownership is recorded even at low-income levels, but becomes significant only when household incomes reach Rs 25,000 or more. Table 8: Vehicle Ownership per 100 Households by Income (Rs/month) - WP 0- 1001- 2001· 3001- 5001· 10001· Over Total 1000 2000 3000 5000 10000 25000 25000 Bicycles 34 15 17 28 38 41 34 33 Motor Cycles 07 04 02 14 11 24 21 14 Cars & Vans 00 01 02 01 04 15 52 09 Percentage of Income Spent on Transport The analysis of expenditure on public transport as a percent of expenditure on transport incurred by three different income groups is given in Table 9. This clearly confirms the earlier trend but also provides information that the income group with less than Rs 3,500/= for monthly incomes are clearly captive to public transport, while this figure falls to around 50% to 60% percent of households when incomes are between Rs 3,500/= to Rs 10,000/=. Land and Urban Policies for Porvety Reduction 125 Table 9: Distribution of HH Income Groups by Expenditure on Public Transport (2000) Income group J Expenditure on Public Transport as a Percentage of Expenditure on Transport 0-20% 40-60% 20-40% 60-70% 70-80% 80-100% Less than Rs 3,500/= 100% Rs 3,500 - Rs 6,000/= 3.7% 5.6% 5.6% 18.5% 11.1% 55.6% Rs 6,500- Rs 10,000/= 9.6% 17.3% 1.9% 3.8% 7.7% 59.6% Expenditure on Public Transport and Income Data from SUS (1999/2000) have been tabulated in Table 10 to show the percentage of household expenditure spent on public transport by income group, for the Western Province. The table shows that the percent of expenditure on transport is below 3 percent for the majority of households, irrespective of their level of income. The higher percentages are to be found among those households with higher incomes. However, it should be pointed out that the vast majority of public transport travel should be undertaken by those in the higher income categories. In this respect it should be noted that since the consideration is by household income and not individual incomes those households with several income-earning members would have a higher income but also a proportionately higher transport cost due to increased travel to work. Table 10 does, however, indicate that the higher percentage expenditure on public transport is concentrated in the middle class households where incomes range between Rs 3,0001= to Rs 25,000/= per month. In the case of those households with incomes less than Rs 3,0001=, less than 2 percent of households incur more than 9 percent of their expenditure on pubic transport and less than 6 percent of the households incur more than 6 percent of expenditure on public transport. The respective values are higher and nearly double in the Western Province. This means that the urban poor appear to spend proportionately more on public transport than the rural poor do. This could be due to difficulties in using alternative modes of transport in urban areas, particularly bicycles; or else it could also be due to longer distances to work and school. Table 1O: Percent Expenditure on Public Transport by Income Group (WP) Percentage Income Group (Rs) Expenditure 0- 1001- 2001- 3001- 5001- 10001- .................................. Over .....................+ Total 1000 2000 3000 5000 10000 25000 25000 0 percent 65.5 69.2 51.2 44.7 35.6 35.0 34.5 41.9 0 to 3 percent 13.8 11.5 9.8 16.0 16.0 26.5 24.1 18.0 3 to 6 percent 13.8 15.4 26.8 23.4 27.0 14.5 27.6 22.0 6 to 9 percent 3.4 3.8 7.3 11.7 8.6 12.8 6.9 9.4 9 to 12 percent 3.4 0.0 4.9 1.1 7.4 6.0 6.9 5.0 12 to 15 percent 0.0 0.0 0.0 3.2 3.7 1.7 0.0 2.2 Over 15 percent 0.0 0.0 0.0 0.0 1.8 3.4 0.0 1.4 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 126 Land and Urban Policies for Porvety Reduction Analysis of the Travel Patterns of the Working Poor in Colombo City The second source of data is from a survey of those identified as the working poor that studies the ability to access work and their residential features such as distance and type of house. The paper also compares the potential to work with the commuting distances. The comparison is based on the relative costs of transport, time of travel, availability of late night travel and social parameters such as type of housing, status of children's education, etc. The survey also investigates the impact on those employed in transport services. The results are particularly significant with respect to three-wheeler (auto-taxi) drivers who are resident in urban settlements and prefer to live close to the city centre which is a focal point of their work and cannot drive their vehicles long distances for the night. They are different for bus crews who usually live a fair distance away from the city centre as they can ride their own bus home for the night. Survey: A total of 164 personal interviews were made of people who were working within the Colombo Municipal City Limits. The questionnaire used for these surveys is given in Annex 1. The survey included location of employment and residence, mode(s) of travel, travel cost and time by each mode, nature of employment, work hours, nature of residence, if transport curtails longer work hours, monthly expenditure, income and household vehicle ownership. The breakdown by employment type is given as follows: I • Security Guards: Mostly earning the minimum legal monthly pay often working double shifts •Parking Wardens: Mostly permanent employees oflocal government • Cleaning Personnel: Government and private sector contracted staff • Labourers: Working on daily wage basis •Traders at Wayside Stalls: Working in fixed areas but self employed. These five groups represent the lowest earning employees in the city. In addition two other groups representing transport-sector workers were also interviewed. These are identified as: •Three Wheeler Drivers: Mostly self-employed auto-rickshaws drivers • Bus Crews: Crews mostly working on daily pay basis for buses owned by private individuals. Distance of Travel & Generalised Cost The mean travel distance and the Generalised Cost of Travel by each group of employees are given in Table 11. The distance is taken as the minimum road based distance for travel computed by the TransPlan traffic model (University ofMoratuwa, 2003). Generalised Cost is computed to represent in addition to the fare or cost of transport, the cost of time, which is calculated at 20% of the income rate. The income rate is calculated by dividing the monthly income stated in the survey form and dividing by the total working hours reported for the month. Land and Urban Policies for Porvety Reduction 127 Table 11: Travel Characteristics of Employees Transport Sector General Employment Employment ..... "' ~ .... :;?;'"' Cl c: Cl Qi Cll "' ~ ~ ~ ........ c: c: ~ "'C .... Cll · - "'C ·c: - "'C ~ c: "' ·- c: c: 0 ::::J 0 ·- "' "'C Cll Cll Cll > ..c: ·- Cll u u ::::J J:: "' "' ::::J "' ~ Cll "' "' Cll - .... Cll .c 10' i= s: "' $: c "' ::::J ll. u ll. "' -' ..... CCI Distance to Work place (km) 7.3 9.2 12.6 8.1 8.6 7.9 14.1 Cost of Travel (Rs/one-way) 10.3 10.5 9.3 11.2 9.1 35.1 4.0 Travel Time (mts/one way) 46.3 52.1 40.9 42.0 33.4 31.1 38.5 Total Generalised Cost/day 31.3 31.2 24.4 30.4 19.0 85.2 18.1 Monthly Income Rs/Month 8,318 6,584 6,914 5,750 10, 111 10.285 12,384 % of Income for Transport 15.1% 9.0% 14.1% 21.1% 7.5% 33.1% 5.8% It is seen from Table 11 that average travel distances between different employment categories vary between 7 .3 kms and 14.1 kms. The travel time varies between 31 minutes and 52 minutes. The travel cost varies between Rs 4.00 for bus crews- who travel for free along with the bus most of the distance and a high of Rs 85.2 for three wheeler operators who have to ride their vehicles to the place of operation. Apart from these extremes demonstrated in the transport sector employment, other employment demonstrates fairly uniform costs and travel times. Figure 3: Variation of% of Income Spent on Commuting to Work and Income 0/o of Income Spent of Transport CV 25% Et: • 20% -.- - -- 0 0 u c.. c Ill - c - ... _ - - - -- - - - >- Rl :i- 15% - .c t- c 0 10% 0 ... :iE c 5% -0 cuQ.. 0 CF- U'I 0% 0 5,000 10,000 15,000 Monthly Income (Rs/Month) Interestingly, however, the relationship between expenditure for transport as a percentage of total income appears to have an inverse relationship with income. As indicated in Table 11, the lowest average income earners who are labourers spend 21.1 % of the their incomes on generalized costs for travel, while the highest income earners who are the wayside traders spend only 7.5% of their incomes on transport. Figure 3 shows 128 Land and Urban Policies for Porvety Reduction this relationship where the lower the average income, the higher is the percentage of their income that is spent on transport. While the bus was the predominate mode of travel for all categories, the higher income earners spent less time to travel to the same distances as they tended to live closer to the main bus routes and the travel times were less. This is intuitively plausible since the higher income groups could afford to live in lands closer to the main roads. Thus the distance from the main bus routes appear to be the primary reason for increase in total travel costs. Land Ownership, Land Prices & Distance to Work The survey by tracing the location of work, alternative housing locations, access and cost of transport etc finds that land prices in suburban areas which are alternative locations for the urban poor to be relocated are usually away from the main transport corridors and are presently poorly served by public transport. The irregular hours that the poor work are not conducive to public transport which usually operates well only during peak periods. The costs of travel to these alternative sites are high - hence, the need to reside in the city. This increases the value of land and also overcrowding in settlement areas which are the only such affordable lands for the poor. In addition this puts pressure on services in urban areas and results in the poor not having adequate equal access to these services which are more freely available in suburban areas. For example, the city has the most popular schools, but the ones attended by the children of the poor are neglected when compared to similar schools in suburban areas. Similarly, the incidence of health and safety problems is higher, as is that for crime and other related activities in these squatter settlements. The house & land ownership of the residences occupied by the interviewees is given in Table 12. It is seen that only 22% of the people were on rented land. While 42.1 % stated that they were occupying legally owned land, 26.8% stated it was government land. The latter are to be considered mostly as squatters on state lands, usually marginal lands in the periphery of the city. The fact that nearly 70% of the people claimed a fixed abode makes them less mobile to seek accommodations closer to their places of residence. This also adds to increased commuting distances and increased transport costs. Table 12: Breakdown of Land & House Ownership Percentage Own Land 42.1% Government 26.8% Rented House & Property 22.0% Other 9.1% This is further reinforced by the evidence that percentage of those who own their own house and property decreases as the distances between residence and work place decreases. This is shown in Table 13, which shows that only 31 % of those living within 5 kms from their places of employment occupy their own houses. This increases sharply to 64.5% when the distance increases to over 10 kms. Land and Urban Policies for Porvety Reduction 129 Table 13: Percentage of Employees who live in Own house & Property with respect to Distance from Work Place Percentage Less than 5 kms 31.0% 5 to 10 kms 50.0% Over 10 kms 64.5% However, the quality of housing appears to fall when employees get closer to their workplaces- i.e. to the centre of the city. As shown in Table 14, those living less than 5 kms from their work places do so in Housing Settlements which have only shared amenities, as opposed to Separate House & Property' or Flats (Apartments). Thus it is clear that while going further away from the city centre has an added advantage, as the quality of housing that can be afforded improves. Table 14: Percentage of Employees who live in Settlements with respect to Distance from Work Place Percentage Less than 5 kms 31.0% 5to10kms 11.1% Over 10 kms 6.5% The value ofland as perceived by most of the interviewees appears to be quite suspect as they seem to have no clear idea of the market value of land. Even those who stated they lived on own land had a poor idea of the actual value. This could also be due to the fact that most of the land which was considered as 'own' is also encroached and not legally owned. Hence the value of exchange of such land is only a fraction of the market price. Moreover, these lands most often located on marginal land used areas such aspn canal banks, marshy areas prone to stagnating water or flooding, under developed localities etc have a depressed market value compared to the better developed and sought after land at equal distance from the city centre. Figure 4: Land Value and Commuting Distance 120000 100000 80000 60000 40000 20000 +--~~~~~~~~~~-~-----< Missing 4.0 5.0 7.0 8.5 11.0 13.0 15.0 20.0 3.0 4.5 6.0 8.0 9.0 12.0 15.0 17 .o 27 .o Distance Travelled (km) 130 Land and Urban Policies for Porvety Reduction However Figure 4, clearly shows that even the perceived land value has a correlation with the distance from the centre. It also shows the lower land value in the city centre itself, which due to the commercial and wholesale trade environment has lower market prices. The highest prices are distances between 2 to 4 kms from the city centre. According to Figure 4, the value of land drops to about l/3rd the cost at distances of 20 kms. Figure 5: Land Value and Generalised Cost of Commuting 160000 ~---~~~-----------, 140000 ,~ ~ .,; es 4J) ::I ~ 120000 100000 I "C 80000 c: +:> Average Population Density Gradient Population Density Gradient distance per 1991 2000 Gross Density Population 2000 person to City Area (km 2) 2000 city center Constant Density Adjusted R2 Constant Density Adjusted Rz (people/ha) (km) Gradient Gradient Sao Paulo 1,525 68.4 10,434,252 14.1 5.30 -0.073 5.04 -0.049 35.7 19.8 (Prob JTJ > tJ) (0,0000) (0,0000) (0,0000) (0,0000) 24.5 .Rio de Janeiro 1,225 78.0 5,857,904 19.9 5.13 -0.040 5.00 -0.029 15.5 (Prob JTJ > tJ) (0,0000) (0,0000) (0,0000) (0,0000) 27.6 Fortaleza 330 64.9 2, 141.402 8.0 5.33 -0.166 5.14 -0.108 4.8 17.9 (Prob JTJ > ti) (0,0000) (0,0000) (0,0000) (0,0000) 1.422,905 26.7 Recife 218 65.3 6.8 4.89 -0.098 4.80 -0.071 2.3 (Prob JTJ > ti) (0,0000) (0,0196) (0,0000) (0,0767) 52.7 Salvador 279 87.4 2,436,390 8.1 5.39 -0.146 5.29 -0.100 40.7 19.4 (Prob JTJ > tJ) (0,0000) (0,0000) (0,0000) (0,0000) 1,587,315 12.0 Curitiba 435 36.5 7.6 4.52 -0.191 4.32 -0.134 34.7 (Prob JTJ > ti) (0,0000) (0,0000) (0,0000) (0,0000) 32.8 1,360,590 7.7 -0.187 5.06 -0.168 Porto Alegre 375 28.8 5.11 42.0 (Prob JTJ > tJ) (0,0000) (0,0000) (0,0000) (0,0000) ,.... !lJ :J c.. 2,238,526 4.73 -0.052 !lJ Belo Horizonte 331 67.6 7.2 4.80 -0.082 :J 6.0 c.. (Prob JTJ > tJ) (0,0000) (0,0017) (0,0000) (0,0211) c:: c- !lJ 1,093,007 3.97 -0.150 :J Goiania 385 28.6 6.7 4.39 -0.203 "C 11.8 2.. (Prob JTJ > tJ) (0,0000) (0,0000) (0,0000) (0,0013) ;;· iii" 2,051, 146 3.09 -0.003 "' ..... Brasilia 5,822 3.5 20.1 (R 2 = 0.0) ... 0 (Prob JTJ > tJ) (0,0000) (0,0000) "C ... 0 < Cl) r+ Source: Census data and the authors' calculations. '< ::0 Cl) c.. c: n ~. 0 :J Land and Urban Policies for Porvety Reduction 145 Figure 1: Population Density Gradients. SAO PAULO 1991 SAO PAULO 2000 • 250 200 • •• -·------------- ... 200 .---------------.---- - --------------- • • • .. •• • • • .!!! ~ 150 c "' "' c ••• .. •- - - - '•----•-- ...•*••• • ... .. --.---- •• •• ---- -- \+ • ------- -- ---- ---------- ----- "' :E "' "' 100 c "' c 150 -- ••• • • • .. .. --~--·--- ·-·-------~.--------- ---,-,-...-.- • t + . . . ----.--.---- - . • • • ------ --- -- -- ------- ---- -- -------- - --- -- -- -- --- - - - . '· . • --------------- -------- 100 -- +"'1 • • • :·~~ - ~--.-.---~-- - -- -- ----------------------- ~. •••• ~ 50 ___ • , --·----'- ·------•• ___ -- - • • ••• • •• • 50 ------~----:;-~-- .~t. ~--:- 0 -.f------.------,-·------.--•---.--+-----1 0 10 20 30 40 50 0 10 20 30 40 50 Distance to the city center Distance to the city center RIO DE JANEIRO 1991 RIO DE JANEIRO 2000 500 450 450 -----------------------------------·------ 400 ---.---.-·------------- ------- ------ • ----- 400 350 .. ·-- • ---------- ---------- ------- • 350 300 • • 300 "' .!!! :!::: :fl· 250 :.;::; 250 'iii "' c c 200 "' 200 c "' c 150 150 100 100 50 50 0 0 10 20 30 40 50 60 Distance to the city center Distance to the city center FORTALEZA 1991 FORTALEZA 2000 350 ~-------------------------. 500 ~--------------------~ • 300 -- - - -- -- -.-- -- -- ---- -- -- ---- - -- -- -- --- - ---- -- ---- ---- - 450 400 250 350 •• • "' ~ 'iii c "' 200 c 150 - --- - ---·-- • .,. ---··- -----.•--+---•-----'----- ----- ----- ---- ------- -----·--~---------- - - ------- - - ------------------------ ~ 300 'iii 250 c "' 200 c -- • ----------------------------------------------- • • ~, .. 1 • , •• • • .# • • • 150 100 ----.--· •i-;.- --- '¢+- ---- + + ----.---- -------- • 100 50 ___! ___ • ~--. __ !_+_~ • __.____ -·-~·---- -- --- --------- 50 + ~ • •• •• 0 +-------,_:_ .. _+ .....·~·· _._,c__...,_:__ _ ___.,..._ _,.___---1 0 +--------.=---"'--"~.------'--tl'--~,.___----1 0 5 10 15 20 0 5 10 15 20 Distance to the city center ·Distance to the city center 146 Land and Urban Policies for Porvety Reduction RECIFE 1991 RECIFE 2000 350 350 300 . 300 ---. - -- -··- 250 ~ -----------..----- •• • 250 -·----·. -·• - --------·--- rn :! 200 ---- - - ------- -·- -•----- ----•- -- -- ---- • •• • - ----- --- -- - -- -- -- -- -- - -- -- ~ rn 200 • .. -. •• 'iii c c 150 Q) 100 --- • . . ... .. -- •·- --·--- _4'.'! __• - +# • •• -- ,_... --!•----• • • ---------•-- ----- ------ ---------- ------ 'iii c c 150 Q) 100 -·-- ------ • • • •~·· •+"> ... -.-- - -. ·--.-:- . . . . . . .• --~-·---- .. - ·-···· - -=- ··-·---·- ---------·---- ,.. : # •• .: ..•• ..·--·- ~- ... T"~ =··- • • • • 50 • •.-- ------.- • .......---.--- • ---;· --- 50 0 .. • i :·. 0 •• 0 5 10 15 20 0 5 10 15 20 Distance to the city center Distance to the city center SALVADOR 1991 SALVADOR 2000 600 600 500 500 ------ ---- ---- • • • • 400 400 -------·----- • • • - - ---- .. t . rn :! •i ~ rn • • -. ---· •• ;. .. 'iii 300 -· ~·­ 'iii 300 .. c c • • - - - Q) c ••• __ ! __ c Q) • • r • • -- ·-·- • . -•-.--1----.--- . 200 200 -- ·.~ • --.--• • ....... 100 0 • - -·------------------ • 100 0 ----.------ . ... • ~·~-·~-·~:--.-.__~~-·~·~·__.•,.__·~~·~~~"--,~~~~-1 ------ --------- 0 5 10 15 20 25 0 5 10 15 20 25 Distance to the city center Distance to the city center CURITIBA 1991 CURITIBA 2000 120 100 100 • • •• •• • • ...... ,. ...._.. 80 . 80 rn ~ 'iii c Q) 60 ..••• --- _ ~­ rn :! 'iii c Q) 60 ... •• • • • -· # - c c 40 -•---- • • --~----~-+-!---•-----~ -- • --------- -------- ----------- - 40 ·------·- •• • •• -- - --~--- • • • 20 •... • 20 ------·- ---.--.-.. '!. -:-- t --~ • • • • 0 5 10 15 20 25 0 5 10 15 20 25 Distance to the city center Distance to the city center Land and Urban Policies for Porvety Reduction 147 PORTO ALEGRE 1991 PORTO ALEGRE 2000 350 - - - - - - - - - - - - - - - - - - - - - - - . 350 -,-----------------------~ 300 300 250 -- ...---- ---------------- ----- -- ------------------- ---------------- 250 .~ 200 ill 200 • ~ • :e c "' c • • 2: 150 - ---~-------- ---------------- ------------ ---- ---- - ------------ • • ••• 2: 150 ----. -- -- --.------ 100 50 • --:~ ---- -·-~-- ---- ··~ .....'·_.__ ._ - - . 100 50 . t._; ..... __:. --- -•- ---- ----- .. ·---• ... .. . _ ___ •• • • ••• •• • o-1---___i_~-,----__.._.__,~~~~===~~t:-_,._~_____j • 0 5 10 15 20 25 30 0 5 10 15 20 25 30 Distance to the city center Distance to the city center BELO HORIZONTE 1991 BELO HORIZONTE 2000 350 350 ----------------------~ • 300 ·- • 300 ----- - -~---- - --------- -------- ------------ ---- ----- ------- --------- • -----•-- ---------- ------------ ---- --------- -- 250 -.--- 250 - - --------- -- ---- ---- ., "' 200 ·------ ~ 200 :;::; "iii • --- -- - "iii • • ., c c 150 . . . . _ :_ _ _ _. +_!_ •. ., c c 150 ------ -----·- • • ----- ----- ------------ ----- -- -·- -- . .•• . ... ~ 100 50 .. . . + -•.,...-~_L.-- , ' : ._ , ••• ~-~· __ - --- 100 50 ·~.. •••• . . .. •.. . • -------.-~·-------·-----·-·- - • ----#~----- --------------- • ·- •• ___.,. ____ - ---.-.----· - ---- ~ ---- ·--.----- ----- ---------- •• • t• ••• 0 • 0 +------·-·~·---·-·--·~·-·-•___.o-------~ 0 5 10 15 20 0 5 10 15 20 Distance to the city center Distance to the city center GOIANIA 1991 GOIANIA 2000 140 120 120 100 • 100 •·- ---. 80 -----.-•-------------·---- ---.-- ------- -------------------------- • • •• ., "' • • • "' .!!! ~ 80 •• +,: :e 60 • ·--•-•-----!~- ----• • ., "' • • ••• • • ., "' c 60 --- _t __ ------·· ---- c • c .·---..--.-·----.---- c 40 ----- -· -:- -· •---·---- •• --- :. • .'•• -· .• - .. 40 ---·----- ------------------------------ 20 ---·-·• •• -·--·------ •• • t - - --.-- 20 • • ... -.--- • ••• • • -.--•- • • • • . 0 0 +-----·--·------•_..,_.---"!~·~·----~·-------1 0 5 10 15 20 0 5 10 15 20 Distance to the city center Distance to the city center 148 Land and Urban Policies for Porvety Reduction BRASILIA 2000 200.00 180.00 160.00 140.00 .-• "' 120.00 ·--- ----.----+-.• - "' • ..• . :;::; "iii 100.00 c "' c 80.00 + • • ---- - --- 60.00 + 40.00 • + • _____ • __.;_ 20.00 •.. - -·- + 0.00 .. + • • +--~>-r-~-""--___.,c*"---"~~-L----~ + + • 0 10 20 30 40 50 Distance to the city center Actually, the pattern of decreasing density with distance from the CBD found in the Brazilian cities of our sample is similar to what researchers have observed in most ci~ies in the world. We note that in 2000 the most centralized cities were Porto Alegre and Goiania, in effect, they presented the highest density gradients among the cities analyzed. These cities have similar area - around 380 km 2 - and a population between 1 and 1.4 million. The city of Recife is comparatively smaller than the former cities in terms of its area, but presented a more decentralized distribution of its population since its population density gradient is smaller. This is probably due to the geography of the city, cut by the canals of the Capibaribe River near the center of the city, increasing the occupation of areas further from the city center. Note that the gross density of Recife's area is greater than the densities of Porto Alegre and Goiania, reaching a value more than twice the gross densities of both those cities. This is to be expected since the Recife's area is smaller than the others. However, though smaller size and larger population can mean a more compact city, the Density Gradient found for these three cities shows that Recife, comparatively, has a more dispersed distribution of its population across its area. Despite this, Recife is the city that has the lowest average distance per person to the city center, of about 6.8 km. Curitiba is the third most centralized city in our sample, having a few more than 1.5 million people distributed in 434 km 2 area and gross density is of about 36.5 persons per hectare. This is nearly half the density of Fortaleza, a city that has 2.1 million people distributed in a smaller area, about 330 km 2 • Although Fortaleza is a more compact city, with smaller area and bigger population, Curitiba is more centralized than Fortaleza, as shown by its higher density gradient. The examples of Curitiba and Recife are important points because they indicate that density gradient does not depend on city size or population, but on the pattern of the population distribution in the city area. In general terms, the pattern of population distribution in Curitiba results in a lower average distance per person from the city center than in Fortaleza, which results in lower transport costs and higher efficiency in urban services supply. Indeed, the average distance per person in Curitiba is 7.6 km, while in Fortaleza it is 8.0 km per person. Rio de Janeiro and Sao Paulo, the largest cities of our sample, presented high average distances per person to Land and Urban Policies for Porvety Reduction 149 the city center, of 19 .9 km and 14. l km, respectively. However, Brasilia presented the highest value of our sample, of 20. l km per person reflecting the dispersion of its urban areas across the territory of Federal District (DF). In 2000, the most decentralized cities of this group were Bras1lia, Rio de Janeiro, Belo Horizonte and Sao Paulo, the largest cities in our sample, except Brasllia. Despite the reliance of the density gradient on the distribution pattern of the population in the built-up area, big cities, in general, tend to be more decentralized than the small ones. This can be attributed to the rise of agglomerative ec?nomies in new areas of the city and subcenters formation that share with the CBD the power to polarize occupation. This reason cannot be attributed to Brasilia's case, since its poly-nuclear spatial arrangement has no relationship with subcenter formation, as we will see later. Sao Paulo, the largest city in Brazil is more centralized than Rio de Janeiro, a smaller and denser city. This reflects the specific topographic conditions of Rio de Janeiro's site and the attractiveness of the sea coast as a desirable location for housing, retail and services, a factor that reduces the attractiveness of central areas. Both cities have several subcenters, but the special conditions of the Rio's site yield a more decentralized city. Belo Horizonte, in spite of its rough topography, similar to the Rio de Janeiro's, although slightly less dramatic, is more centralized. The decentralization of the population in Belo Horizonte between 1991 and 2000 was stronger than the decentralization in larger cities such as Sao Paulo and Rio de Janeiro. The higher population growth rate of Belo Horizonte in the period can explain this fact. Note that the urban growth in Brazilian cities is more intensively absorbed by peripheral areas. Brasilia, which will be analyzed separately, presents a density gradient curve practically flat, with a slight negative slope. But if some aspects of its spatial structure are considered separately its density gradient inverts showing a positive slope. The results shown in Table 4 reveal that while the estimated density gradient found for Brasilia is negatively sloped it has no statistical significance. The R 2 also demonstrates that the endogenous variable deployed to explain the density distribution of the city, e.g., the distance to the central areas, does not effectively explain the pattern of population distribution across the city. This pattern is a product of the state land market arbitrage combined with strong land use controls, which yield high densities far from the central area of the city, while the city center presents low densities. Table 4 shows that most of the cities analyzed present an average distance to the city center per person around 7.5 km, except for the largest cities of our sample and Brasilia. This value is about the same of the median distance per person found in cities in other countries. For example, in 1993 Paris had a median distance per person of 7 km and Moscow had a median distance of about 10 km (Bertaud and Renaud, 1997). The larger median found for Moscow reflects some specific features of its spatial structure, as the misallocation 6fland near the central areas that exacerbate the increase of the median distance per person to the city center. Figure 2 show that within 5 to 9 km radius from the city center most Brazilian cities accumulate the major share of their population. This explains average distance per person to the city center of about 7.5 km, in median, found in seven cities of our sample. The exceptions, Sao Paulo, Rio de Janeiro and Brasflia presented different pattern of cumulative population across their urban areas: Brasilia, due to its spatial dispersion and the others due to the scale of their urban areas. 150 Land and Urban Policies for Porvety Reduction Figure 2: Population in the built-up areas of 10 Brazilian cities. Distribution of Population Distribution of Population Sao Paulo 2000 Rio de Janeiro 2000 1,eoo.ooo~-------------------- 600,000 1,400,000 500,000 1,200,000 j 1,000,000 c 0 400,000 l 800,000 600,000 i J?. 300,000 200,000 400,000 --- 100,000 200,000 0 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 0 3 5 7 911131517192123252729313335373941434547495153 Distance from center city Distance from center city Distribution of Population Distribution of Population Fortaleza 2000 Recife 2000 450,000~-----------------------, 4so,ooo~--------------------- 400,000 400,ooo 350,000 350,000 5 aoo,ooo ~ 250,000 j :~~:~~~ [ 200,000 a 200.000 j_ 150,000 ~ 150,000 100,000 100,000 50,000 50,000 0 11 13 15 17 11 13 15 Distance from center city Distance from center city Distribution of Population Distribution of Population Salvador 2000 Curitiba 2000 500,000 soo,ooo~---------------------~ 450,000 450,000 400,000 400,000 c 350,000 ~~~:~~~ ~ :~~:~~~ E ~ 250,000 i 200,000 g. 200,000 150,000 c.. 150,000 100,000 100,000 50,000 50,000 0 11 13 15 17 19 21 11 13 15 17 19 21 23 Distance from center city Distance from center city Distribution of Population Distribution of Population Porto Alegre 2000 Belo Horizonte 2000 350,000~---------------------- 600,000 300,000 500,000 i l 250,000 200,000 150,000 il 400,000 300,000 200,000 100,000 50,000 11 13 15 17 19 21 23 25 _5 11 13 15 Distance from center city Distance from center city Distribution of Population Distribution of Population Goiania 2000 Brasilia 2000 250,000~---------------------, 200,000 1150,000 f 100,000 50,000 11 13 15 17 19 03579ttuwnm~a•v~~m•u•Ma Distance from center city Distance from center city Land and Urban Policies for Pqrvety Reduction 151 Indeed, the share of the population that lives within a 5 to 9 kilometers radius from the city center ranges from 52% to 55% of the total in Salvador, Belo Horizonte, Fortaleza, Porto Alegre and Goiania. In Recife the share of the population that lives at that area corresponds to 73% of the total. In Sao Paulo and Rio de Janeiro, that population is about 20% of the total, while in Brasilia the population that lives at the same distance from the city center is only 15% of its total. In general, all values measured for Brasilia are close to the values reached by the cities that are the core of the largest cities of Brazil, despite the fact that it has only one fifth of the population of Sao Paulo and less than half of the population of Rio de Janeiro. · Brasilia and its spatial pattern of population distribution - Is the Functional City an efficient city? Brasilia is a sui generis city. Its original urban plan, the Plano Piloto (Master Plan), was chosen in a public contest in 1957 and its construction took only three years, being inaugurated in 1960. Its plan, designed by the architect Lucio Costa, is tied to the principles of modernism, an architectural movementthat preached a particular formal model to conceive cities. Among these principles was the idea that cities of the modern industrial age shoul :l be organized according to functional areas, connected by transportation axes that would allow fast linkage among them. The principles to design the "Functional City'' were compiled during the IV International Congress of Modern Architecture (CIAM), held in 1933, in the document named Athens Charter. This document pointed out the four functions that were considered the key issues for the Functional City: to reside, to work, to recreate and to circulate. In order to build the new capital, the government dispossessed great extensions of farm land, placing in state's hands most of the land in the Federal District. Thus, public administration has been the main agent allocating land for housing, agricultural, industrial and commercial activities. In addition, land use within the Plano Piloto must obey the guidelines traced by Lu~io Costa's project and does not allow for changes. Since its inauguration Brasilia attracted an intense migratory flow of people from the States of Goias and Minas Gerais, and from the poorest areas of the north and northeast regions of Brazil. Even though government owned most land. it was unable to respond effectively to the housing demand growth, especially from the poorest groups. Between 1960 and 1980, for example, Brasilia presented a population growth rate of about 11.2% per year, passing from 130, 796 inhabitants to more than 1.17 million in 1980. As a result, a great number of irregular settlements formed near to the Plano Piloto, even though some of them started from construction camps that were not removed after construction works had ended. In order to respond to the increase of housing demand, the government constructed several residential settlements at the periphery of the Plano Piloto to provide low cost housing not only for the people that inhabited the slums, but also for its administrative workers. Often, the addition of new residential areas, especially those not anticipated by the original plan, was placed far from the Plano Piloto. The locations assigned to settle the new urban areas were determined by the several territorial plans formulated over the years to control urban occupation within the Federal District. These plans responded to 2 main objectives: protect the Plano Piloto from changes caused by market constraints and preserve the basin of the Paranoa Lake in which the Plano Piloto is located from densification. The efforts to preserve Brasilia's Plano Piloto culminated in 1987 when it was assigned as World Cultural Heritage Site by UNESCO. With little more than 37 years of existence Brasilia had its principles of spatial 152 Land and Urban Policies for Porvety Reduction ~ organization preserved as an inportant cultural site, the only contemporary city to receive this title. & such, the principles of conception of Brasilia were preserved according to the 4 scales on the basis of in which the city organb:.~tioh was conceived: the Bucolic Scale, formed by the green and empty spaces that surround functional ar~as; the Gregarious Scale, formed by the spaces that are set to work activities; the Monumental Scale, the spaces that form the administrative core of the city, and the Residential Scale, the housing areas of the Plano Piloto. According to these guidelines, the relationship between built and empty spaces should be protected from using opened spaces for new developments. In addition, the original height of the buildings must be kept unchanged, and land recycling via demolition and re-building of new and higher structures in response to the increase of land price is not allowed. During the 1980's and 1990's, Brazil's economic outlook changed. Due to accelerating inflation, rising internal and external deficits and economic recession, easy public financing for housing ended. This conjuncture paired with rigid regulation of land use decreased the supply of housing in Brasilia, for low income and for the middle classes as well. The robust demand for housing in Brasilia has stimulated the emergence of a private land market that trades the undeveloped private land not yet expropriated. In addition, an illegal land market was encouraged on public land. Since then, many developments of detached houses in private and public land have occurred in the irregular and illegal land market not only for low-income people, but also for middle-income families. This so-called "horizontal condominiums" problem has been the subject of extensive discussion, several studies and an evolving set of policies. Figure 3: Urban structure of Brasilia. Informal and irregular settlements D Heritage site limits 0 5 10 15 20Km -Low income Environmental areas ~~~~iiiiiiiiiil li!iil Middle-class I :, :/Lakes Source: State Secretary of Urban Development and Housing of Federal District (SEDUH-DF), 2004. Land and Urban Policies for Porvety Reduction 153 The sketch presented above summarizes the historic background of the spatial formation of Brasilia, that is shown in Figure 3. Note how the spatial structure of Brasilia is fragmented, reflecting, among other things, the official control of land supply. In addition, very often the spaces between the urban agglomerations that form the urban mosaic of Brasilia have environmental restriction against occupation. Nonetheless,, , these areas have constituted a stock of undeveloped serviced land that distorts land prices in the overall city and are, in addition, targets of informal occupation. The urban areas of Brasilia fall within a 43 kilometers radius but the built area corresponds only to 598 km2, of about 10% of the area defined by that radius. Hence, if we consider only the built-up area, the effective density rises from 3.5 to 33.5 persons per hectare. The low density and great distances found in the Federal District result in large urban services delivery costs that are paid by the whole society. Nevertheless, these costs are disproportionately borne by the poor people who live in the most distant areas of the city. While the population of Brasilia sprawls over a 43 kilometers radius, job opportunities are heavily concentrated in the central areas. Accordingly to Ministry of Labor (MTE) data, about 70% of the formal jobs in Brasilia are concentrated within the heritage perimeter. Hence, daily commuting distances and costs are quite significant, since most of the population lives far from the city center. As a consequence, public transportation in Brasilia is one of the most inefficient in Brazil. While the median of passenger transported by public transportation is between 1.8 and 2.5 per kilometer in cities such as Recife, Curitiba and Porto Alegre, in Brasilia it is less than one, reaching only 0.82, according to Transportation Ministry data. This indicates that the system in Brasilia is more costly to operate, reflecting the high public transportation fares. In addition, with an average distance per person of about 20.1 kilometers, the opportunity cost of the time involved in commuting is very high, even for poor people. Perhaps, this may stimulate the use of cars. In Brasilia there are 3.4 people per vehicle, an automobile-intensive ratio that, even though higher than other cities, such as Porto Alegre (3.0), Curitiba (2.4) or Goiania (2.4) is lower than Brazilian median rate, of about 5.8. Histograms derived from Figure 2 show that in Brasilia most of the population is located between 19 and 25 kilometers from the city center, while in the others cities in our sample most people are between 5 and 9 kilometers from central areas. In fact, the plots shown in Figure 4 present the cumulative population across the urban area in those cities. They show that, while Brasilia ~ontains 278,364 people within a 7 km radius of the city center, cities such as Belo Horizonte, Salvador and Fortaleza, also with a total population above 2 million, aggregate 1.2 million people in the first two cities and 0.9 million in the last, at the same distance. The 7 km radius in Brasilia includes the residential areas within the heritage limits. If the radius broadens to 9 km, Brasilia presents a cumulative population of 312,487, what corresponds only to 15.6% of its total, while most of the other cities present cumulative population that varies from 60.5% in Fortaleza, with 1.2 million people, to 81 % in Recife, with 1.15 million people. 154 Land and Urban Policies for Porvety Reduction Figure 4: Cumulative population from the city center in 2000. Cumulative Population from city center (Brasilia and selected cities) c: 0 ~ ~ii' o"" a.. ~ Ill,,, >:I ·- 0 ~ :5 ::i~ E :I u 0 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 Distance from city center -a- Belo Horizonte -.-Fortaleza _.._Curitiba -+-Brasilia -a- Goiania -tr- Recife -a- Porto Alegre -.+-Salvador Cumulative Population (%) from city center (Brasilia and selected cities) c: 0 ~ :i c. 0 n.. ~ Cl.I e..... > ~ :; E ::::J (J 0 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 Distance from city center (km) - - Belo Horizonte _,,_ Fortaleza __._ Curitiba -+-Brasilia __,,__Goiania __,,_Recife --0- Porto Alegre -+-Salvador Distances in Brasilia are only comparable to those in Rio de Janeiro and Sao Paulo, even though having much less people. Figure 5 shows that within the same radius 9 km from the center formerly mentioned, while Brasilia has 312,487 people, Sao Paulo has 2.9 million and Rio de Janeiro 1.3 million people. Comparing Brasilia with these cities is important to demonstrate how the spatial model of Brasilia increases land consumption; in a country with serious limitations on access to serviced land by poor people this spatial pattern decreases access of low-income groups to the benefits of urbanization. On the other hand, it is not true that cities that use more intensively and efficiently their urban land yield necessarily a more equitable urbanization. In general, it is not true and the spatial segregation of low-income population recurrs in the Brazilian urbanization model, in one way or another. Land and Urban Policies for Porvety Reduction 155 Figure 5: Cumulative population from the city center in 2000. Cumulative Population from city center (Brasilia, Rio de Janeiro and Sao Paulo) 11 .ooo -,.-,-,..,...-,-,-, i..,...i ,...,....,-.,...,...,..,....,....,..,...1...,-,..-,,I,I..,...I .,.,_:! .,...I 1..,...1.,....1!"1-.,.-,--!1"1..,....,...,:,....,.,'"T"'1.,....,,...,.,..,...1,....,....,...,...,...,[...,--c-,,...,...., ,...,I' 10.000 : I ,) I_, T I I 9.000 ! ,,Vif! I !.-~ :;:::; 40,0% ca :; 30,0% E ::s 20,0% 0 10,0% 0,0% 0 3 5 7 9 11131517192123252729313335373941434547495153 Distance from city center (km) /__.___Brasilia --m- Rio de Janeiro _,,__Sao Paulo I Figure 1 showed that the Density Gradient of Brasilia is negatively sloped, as predicted by the theory, but practically flat. To deepen our analysis, we made an additional set of estimates of the Population Density Gradient for Brasilia considering what we called as "Formal City" and ''Appropriated City". In the first we considered 32 areas that form the regular and official city. In these areas we included the Plano Piloto and the settlements that were developed by local authorities. The second set of regression calculations considered all the areas of the city, including those developed by informal land markets. The areas originating from infor- mal land markets included low-income and middle-class settlements that were developed in private and public land, near the main transportation corridors and near some consolidated urban areas. The results are shown in Figure 6 and respective Tables. 156 Land and Urban Policies for Porvety Reduction Figure 6: Population Density Gradients for Brasflia in 2000. THE FORMAL CITY THE APPROPRIATED CITY 200.00 ~---------------~ 200.00 - - - - - - - - - - - - - - - - - 180.00 - - - - - -- ----- --- -- - - ------ ------ ---- 180.00 --------------------------------------- • • 160.00 140.00 160.00 140.00 -. lll 120.00 ------.-- ---.---- · - - · - - - - - - - - - - -- ------- .. -- "' 120.00 -- - - -- -. --.-- •• --- - - -- -- - ----- --- -- - - - -- - -- -------- - .... :;::; 'iii 100.00 c c: "' 80.00 60.00 ----- -- ----- - • ---------------.- ·--- .-- -.-- • .. .. - - - - - - - ----- -----.+- ----~ -- • -- ----- · - .. ______ - - ---- .. .. _ •- --; -- - -- -·--·---- -- :! ·;;; c: "' c 100.00 80.00 60.00 -- • -------~-- -------- -- --- • - ·----·--------•---------------------- - -. --· • - -.---------------- -----~---- ..•------- ~----·- ------ - - -- --··----------- -- 40.00 40.00 ---- ~--------- -- . • ... -----•- +!____ ------ -~- - 20.00 0.00 •• -+-----~---=---"'----r-----,.------j • 20.00 • ......•... • :.. _ 0.00 -1-----<~-"'L--"-..___....,--_ _.,,~·--l'·'---=·-----l • 0 10 20 30 40 0 10 20 30 40 50 Distance to the city center Distance to the city center Constant 2.66 Constant 3.99 (Prob>/ t /) (0.0000) (Prob>/ t /) (0.0000) Density Gradient +0.0486 Density Gradient -0.0033 (Prob>/ t /) (0.0653) (Prob>/ t /) (0.8479) R2 10.9 R2 0.0 Note that in the first regression, e.g., the Formal City, the Density Gradient curve does not fit the data very well, since the R 2 shows that only 10.9% of the total variation in density is accounted by variation in the distance from the city center. In the Formal City, the spatial distribution of population was not driven by the proximity to the jobs location. On the contrary, the strategy adopted by the authorities of Brasilia was to settle people far from the Plano Piloto, according to planning guidelines concerned with environmental and urban preservation. Hence, the positively sloped density gradient curve reveals the pattern of population distribution promoted by the technocratic and, some would say, elitist planning practice underlying the housing policies implemented over the course of time. Nevertheless, distance to the city center is not a good predictor of population density. In order to correct the model specification it should incorporate other variables, such as proximity to other employment areas, some amenities and, in special way, the rigor of land-use controls. This is dear when we observe the outcome of the second regression, for the Appropriated City, where the Density Gradient curve has a slight negative slope. In this case the R 2 is zero, which signifies that the distance to the city center cannot explain the variation in the dependent variable, the population density. In other words, even though the pattern found corresponds to theoretical predictions of decreasing population density with distance from the city center, the regression line is practically horizontal. In this case, the coefficient of the regression is zero and the values of the densities can not be predicted by variations in distance to the CBD. Hence, in any case the variation in distance was not a good predictor of densities across urban area of Brasilia. However, the reversion of the slope of the density gradient when informal settlements are included in the regression calculations demonstrates that these settlements soften the trend of isolating people from the Land and Urban Policies for Porvety Reduction 157 central areas. This suggest that land allocated by informal markets gets people closer to central areas, compared with command-and-control land allocation. The bad new is that, even considering informal settlements, poor people continue to live far from the main job locations, as shown in Figure 3. Generally, the maintenance of low densities in the central areas is costly and inefficient from the economic land allocation point of view. The opportunity cost involved in maintaining the well-located land undeveloped instead of allocating it for alternative and more efficient use has pushed housing prices higher across the whole city. The increase of the land and housing prices is noted not only in the central areas, but also in areas far from the central ares developed to respond to the housing demand. Aguas Claras, for example, located distant 20 km from the city center, has experienced a fast rise in its housing prices that has reached 40% during 2004. Moreover, the high housing prices are responsible for pushing the poor population further away from the center of the city. It seems unlikely that the green-space and amenity benefits of the vast undeveloped areas could offset the opportunity cost derived from its fragmented spatial structure. The high land prices found in Brasilia suggest that the city, in comparison with others, probably employs capital inputs intensively, perhaps excessively so, in housing production. This reflects the high price per unit of floor area found in Brasilia, higher than in other cities, It is interesting to note that the high land prices that trigger construction of vertical structures, i.e., causing higher densities within the central areas in other cities, in Brasilia operates within structures. Since the height of buildings is limited due to constraints imposed by land-use controls, developers have produced more and smaller housing units per block and per unit ofland. One cost of constraining housing supply in central areas is that households live in increasingly smaller spaces not only in apartment blocks, but also in commercial buildings, where offices are used as housing. Different spatial arrangements impose different pattern of costs and benefits on the city as a whole. The maintenance of undeveloped land nearby and within the central areas via strong land regulations should be evaluated according to cost and benefit considerations and not only according to formal aspects. While it is unquestionably true that green spaces in the core of the urban area are positive for quality of life, due consideration should be given to its cost, since this quality of life is disproportionately appropriated by a small segment of the population. These costs are paid by everyone, including low-income families that live on the fringe of the city. However, these people do not take advantage of the possible benefits of living at great distances from the city center, such as larger houses and a better environment, a normal compensation for the greater costs involving long commuting distances. Conclusion Even though the differences that distinguish the birth and the growth of cities located in different regions of Brazil, including the country's presents great regional disparities, the population distribution across their urban areas is similar to cities worldwide. According to theory, urban population densities decline from the city center toward the outer limits of the city due to high land prices at more accessible sites that trigger the increase of built floor area per land area. By estimating the population density gradients for ten Brazilian cities this pattern was found for nine of them, except forf Brasilia, which the expected pattern was inverted. The results for the population density gradient found for Brasilia, positively sloped when considered only the "Formal City" and slightly negative if formal and informal settlements are considered, are emblematic of the 158 Land and Urban Policies for Porvety Reduction effects ofland allocation commanded by bureaucratic decision and assisted via strong controls over land use. It is worthy noting that Brasilia shares a positive sloped PDG with other cities where land allocation is assigned via administrative processes and assisted by strong land-use controls, such as cities from transition countries such as Moscow (Bertaud, A. and Renaud, B. 1997) or cities where government has implemented apartheid policies in the past such as Johannesburg or Capetown (Bertaud, A. and Malpezzi, S., 2003). Behind the Population Density Gradients found for Brasilia there is a city whose population is dispersed in an area approximately twice that of cities with approximately the same total population. The median distance per person of above 20 kilometers found in Brasilia, for example, shows the dramatic distribution of much of its population at the outer limits of the city. The outstanding urban design of Plano Piloto and the spatial arrangement of the city as a whole have resulted in high land prices and high intra-urban, out-of-pocket and time travel costs. The supply of primary infra- structure also involves high costs due to large distances involved, exacerbated by extensive empty spaces between urban areas. Probably, the costs derived from that spatial arrangement hinder the competitiveness of the city, private investments and development in general. The importance of the study of density gradient patterns lies in the help it provides in formulating efficient public policies in urban areas. In sprawling cities characterized by low densities, exclusionary land uses and bedroom peripheral areas, the supply of services and infrastructure is costly compared to more compact cities. In addition, it generates long commutes that, in turn, increase air pollution and traffic jams. On the other hand, too dense cities may constrain infrastructure systems exposing them to diseconomies entailed by externalities like air pollution and traffic jams. Nevertheless, it is possible to have denser areas that balance a mix of land uses in places accessible by foot and where open spaces have interesting qualities that stimulate social interaction. In addition, the long- term study of densities, combined with the study ofland use, can provide elements for understanding land market dynamics and the development of urban land-value, which in turn is tied to the social stratification of the city space. References Alonso, W, 1964. Location and Land Use. Harvard University Press. Anas, A., Arnott, R. and Small, K., 1998. Urban Spatial Structure. Journal of Economic Literature, Nashville, v. XXXVI, n. 32, p. 1426-1464, September. Berry, B.]. L. and Horton, F. E., 1970. Geographic Perspectives on Urban Systems. Englewood Cliffs, Prentice- Hall, 1970, Ch. 9, "The Urban Envelope: Patterns and Dynamics of Population Density'', p276-305. Bertaud, A. and Renaud, B., 1997. Socialist Cities Without Land Markets, in Journal of Urban Economics 41, 137-151. Bertaud, A. and Malpezzi, S." 2003. The Spatial Distribution ofPopulation in 48 World Cities: Implications for Economies in Transitions. The Center for Urban Land Economics Research, The University ofWisconsin-Madison. I Clark, C.,.1951. Urban Population Densities, Journal of the Royal Statistical Society, 114., pp. 375-386. ' ,I Land and Urban Policies for Porvety Reduction 159 DiPasquale, D. and Wheaton, W C., 1996. Urban Economics and Real Estate Markets. 1 ed. Englewood Cliffs: Prentice Hall. 378 p. Malpezzi, S. and Guo, W, 2001. Measuring 'Sprawl:· Alternative Measures ofUrban Form in US. Metropolitan Areas. The Center for Urban Land Economics Research. Revised, January 15. Mieszkowski, P. and Mills, E. S., 1993. The Causes of Metropolitan Suburbanization. Journal of Economic Perspectives, Vol. 7, Number 3. Pages 135-147. Summer. Mills, E. S., 1972. Studies in the Structure ofthe Urban Economy. Ch. 3, Population and Employment Density Functions, pag. 34-58. Baltimore, The Johns Hopkins Press. Mills, E. S. and Hamilton, B. W, 1994. Urban Economics. 5 ed. New York: Addison-Wesley Educational Publishers. 481 p. Mills, E. S. and Lubuele, L. S., 1995. Projecting growth ofmetropolitan areas. Journal of Urban Economics 37, 344-360. Muth, R. F. 1969. Cities and Housing. University of Chicago Press. O'Sullivan, A., 1996. Urban Economics. 3rd ed. Chicago: Homewood III, Irwin Series. 73lp. Villa